University of Southern California Dissertations and Theses

All-optical signal processing toward reconfigurable optical networks

(USC Thesis Other)

All-optical signal processing toward reconfigurable optical networks

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content All-Optical Signal Processing Toward Reconfigurable Optical Networks by Zahra Bakhtiari A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree Copyright 2015 DOCTOR OF PHILOSOPHY (ELECTRICAL ENGINEERING) August 2015 Zahra Bakhtiari Dedication To my amazing parents, Tooba Torabian and Majid Bakhtiari my brothers, Mehdi and Mohammad for their everlasting love and support. ii Acknowledgements I would like to express my deepest gratitude to my academic advisor and dissertation committee chair, Prof. Alexander A. Sawchuk, for his support and mentorship throughout my Ph.D. program. His knowledge, positive attitude, patience and encouragement truly helped me through my graduate study. He is indeed a role model to students as an academic adviser and I hope I can eventually develop the same characteristics throughout my professional career . I would also like to thank Prof. Alan E. Willner, Prof. Andrea Armani, Prof. Stephan Haas, Prof. Andreas F. Molisch and Prof. B. Keith Jenkins for their guidance and support during my qualifying examination and Ph.D. defense. Finally, I would like to thank all my family and friends. Their unconditional love and support are the greatest blessing in my life. iii Dedication Acknowledgements List of Figures List of Tables List of Acronyms Abstract Table of Contents ii iii vii xviii xix xxiii Chapter 1: Introduction.................................................................................................................... 1 Chapter 2: Basic and Advanced Optical Modulation Format Generation and Detection........................................................................................................................... 7 2.1 Why Advanced Modulation Formats? ............................................... 7 2.2 Electro-Absorption Modulators (EAMs) ........................................... 12 2.3 Mach-Zehnder Modulators (MZMs) ................................................. 13 2.4 On-Off Keying (OOK) ....................................................................... 15 2.5 Differential Phase Shift Keying (DPSK) ........................................... 16 2.6 Differential Quadrature Phase Shift Keying (DQPSK) ..................... 19 2.7 16-Quadrature Amplitude Modulation (16-QAM) ............................ 21 2.8 Coherent Receiver Hardware and Algorithms................................... 23 Chapter 3: Cascaded Second-Order Nonlinear Techniques for High Speed Reconfigurable Optical Add-Drop Multiplexing (ROADMs) ................................................................ 27 3.1 Introduction....................................................................................... 27 3 .2 Cascaded Second Harmonic Generation and Difference Frequency Generation ( cSHG/DFG).... ................ ...... ........................ 28 3 .3 Analytical Model of a Periodically Poled Lithium Niobate Waveguide ............................................................................ 29 iv 3.4 Cascaded Sum Frequency Generation and Difference Frequency Generation ( cSFG/DFG) ...... ...... ................. ..... ................. 33 Chapter 4: High Speed Reconfigurable Polarization-Based Optical Multiplexing Technique for QAM Signal Generation ........................................................................................... . 4.1 Introduction ....................................................................................... . 35 35 4.2 Concept.............................................................................................. 36 4.3 Experimental Setup and Results ..... ..... ................. ...... ................ ....... 40 Chapter 5: High Speed Reconfigurable Nonlinearity-Based Optical Multiplexing Technique for QAM Signal Generation ............................................................................................ 47 5.1 Introduction........................................................................................ 47 5.2 Concept.............................................................................................. 48 5 .3 Experimental Setup and Results ........................................................ 52 5.4 Further Discussion: Optical Multiplexing of a 16-QAM and a QPSK into a 64-QAM Signal 58 Chapter 6: High Speed Reconfigurable Nonlinearity-Based Optical Demultiplexing Techniques for QAM Signals Processing ....................................................................... 61 6.1 Introduction ........................................................................................ 62 6.2 Phase Sensitive Amplification Concept.............................................. 63 6.3 QPSK Signal Demultiplexing Concept............................................... 64 6.4 Results and Discussions...................................................................... 66 6.5 All Optical Logic/Arithmetic Gate Implementation for QAM Sub-Channels ........ ........... ........... ........... ............ ........... ............ 68 6.5.1. Concept............................................................................ 68 6.5.2. Experimental Setup .......................................................... 71 6.5.3. Experimental Results and Discussions............................ 72 v 6.5.4. Experimental Results of Logic/Arithmetic Gate Implementation Between Sub-Channels.......................... 75 Chapter 7: All-Optical Phase-Preserving Multilevel Amplitude Regeneration Using Coherent Polarization Mixing .......................................................................................................... 77 7.1 Introduction ........................................................................................ . 78 7 .2 Concept of Multilevel Amplitude Regenerator Based on Polarization Wave Mixing ................................................................. . 7.2.1. Setup .................................................................................. . 7.2.2 Multilevel Amplitude Regeneration in a Back-to-Back Schematic Results and Discussions .................................... . 7.2.3 Multilevel Amplitude Regeneration in a Nonlinear 88 91 92 Transmission Line Schematic Results and Discussions........... 97 7.3 Dispersion Tolerance Enhancement Using Optical Duobinary Detection in an Optimized 20-70 Gbitls NRZ-OOK Transmission ..................................................................................... 102 7.3.1. Results and discussions ................................................... 105 7.4 Concept of Phase-Transparent NOLM-Free Multilevel Amplitude Regeneration Based on Polarization Wave Mixing.......... 109 7.4.1. Setup ................................................................................. . 111 7.4.2 Multilevel Amplitude Regeneration in a Back-to-Back Schematic Results and Discussions ....................................................... . 112 7.4.3 Multilevel Amplitude Regeneration in a Nonlinear Transmission Line Schematic Results and Discussions ......................... 115 Chapter 8: Conclusion......................................................................................................................... 117 References .......................................................................................................................... 120 vi List of Figures Figure 1.1: ......................................................................................................................................... 4 Three main fully optical modules in optical communication systems. Figure 2.1: ......................................................................................................................................... 9 Single-polarization spectral efficiency versus the received SNR per bit. The Shannon limit for a linear. additive white Gaussian noise channel is shown together with the theoretical performance of various square QAM formats (blue circles). assuming Gray-coded symbol mapping and state-of the-art 7% overhead hard-decision FEC. Also shown in (a) are representative experimental results (red squares); numbers indicate QAM constellation sizes. In (b). the Shannon limits for various square QAM constellations are shown. and the effects of constellation shaping. coded modulation. and signal over-filtering are indicated. [65]. Figure 2.2: ......................................................................................................................................... 12 Trade-off between dual-polarization spectral efficiency and transmission reach. showing the nonlinear Shannon limit of [ 46] together with experimentally achieved results (circles). The ellipse indicates a range into which commercial systems might fall. and the asterisk represents Alcatel Lucent' s commercially deployed optical transmission platform [60], [65]. Figure 2.3: ......................................................................................................................................... 13 Transmission functions of: (a) EAMs and (b) MZMs [35]. Figure 2.4: ......................................................................................................................................... 15 Overview of different options in driving an MZM (Black circles: MZM quadrature points) [35]. Figure 2.5: ......................................................................................................................................... 16 Optical spectrum and optical intensity eye diagrams of NRZ-OOK signal [35]. vii Figure 2.6: ......................................................................................................................................... 17 Constellation diagrams of (a): OOK: (b): Differential phase shift keying (DPSK) [35]. Figure 2. 7: ............................................................................................................................. ............ 18 RZ-DPSK transmitter based on two different approaches: (1) Phase modulator. (2) Mach-Zehnder modulator (MZM). Intensity and phase waveforms of PM-based method and MZM-based method are different [35]. Figure 2.8: ......................................................................................................................................... 19 (a):Optical spectrum and optical intensity eye diagrams of NRZ-DPSK signal: (b): Optical spectrum and optical intensity eye diagrams of 33% RZ-DPSK [35]. Figure 2.9: ................................................................................................. ........................................ 19 Constellation diagram of the optical field of differential quadrature phase shift keying (DQPSK). Figure 2.10: ....................................................................................................................................... 20 DQPSK transmitter (In-phase/quadrature (I/Q) modulator): Two MZMs operate as phase modulators and their outputs combine after applying 90° phase shift on one arm. This modulator is known as an I/Q modulator [35]. Figure 2.11: ....................................................................................................................................... 21 (a):Optical spectrum and optical intensity eye diagrams of NRZ-DQPSK signal: (b): Optical spectrum and optical intensity eye diagrams of 50% RZ-DQPSK [35]. Figure 2.12: ....................................................................................................................................... 22 Symbol diagrams of the optical field of 16-Quadrature Amplitude Modulation (16-QAM) constellation. Figure 2.13: ....................................................................................................................................... 23 16-QAM transmitter: including two parallel (In-phase/quadrature (I/Q) modulators and a 6dB attenuator. viii Figure 2.14: ....................................................................................................................................... 24 Setup of the polarization-diversity coherent receiver [28]. Figure 2.15: ....................................................................................................................................... 25 Block diagram of the digital signal processing used within the coherent receiver [28]. Figure 3.1: ......................................................................................................................................... 28 Mechanism of cSHG/DFG processes in a PPLN waveguide: (a) The CW pump wavelength set near the QPM wavelength; (b) The signal wavelength set near the QPM wavelength [61]. Figure 3.2: ......................................................................................................................................... 34 Mechanism of cSFG/DFG processes in a PPLN waveguide: The CW pump and main signal set at the same distance from QPM wavelength. Figure 4.1: ......................................................................................................................................... 37 Concept of a conventional electro-optical method to generate 16-QAM signal. Figure 4.2: ......................................................................................................................................... 38 Concept of the polarization-based QAM generation method. Figure 4.3: ......................................................................................................................................... 39 Concept and principle of (a) vector addition and (b) tunable 16-QAM generation. Figure 4.4: ......................................................................................................................................... 41 Experimental setup. CW: continuous wave: PC: polarization controller: MZM: mach-zehnder modulator: DGD: differential group delay: EAM: electro-absorption modulator: VOA: variable optical attenuator: BPF: band pass filter: LO: local oscillator: BR: balance receiver: ADC: analog to-digital converter. Figure 4.5: ......................................................................................................................................... 41 Experimental Optical spectrum of 40-Gbaud 16-QAM. ix Figure 4.6: ......................................................................................................................................... 42 Experimental results. BER performance of 10. 15. 20-Gbaud 16-QAM signals versus received power. Figure 4. 7: ............................................................................................................................. ............ 42 Experimental results. BER performance of 10,20.40-Gbaud 16-QAM signals after demultiplexing to 10-Gbaud versus received power. Figure 4.8: ......................................................................................................................................... 43 Experimental Constellations and eye diagrams of 40-Gbaud QPSK and 16-QAM after demultiplexing to 10-Gbaud. DEMUX: demultiplexing. Figure 4.9: ............................................................................................................................. ............ 44 Experimental results. BER characteristic of 10.7 and 12-Gbaud 16-QAM signals as a function of received power before and after DGD off sets: (insets) Constellation diagram of 12-Gbaud 16- QAM signals with and without DGD offset. Figure 4.10: ....................................................................................................................................... 44 Experimental power penalty characteristic of a 20-Gbaud 16-QAM signal versus DGD offset. Figure 5.1: ......................................................................................................................................... 49 Block diagram of optical multiplexing of 4 OOK channels into a 16-QAM signal. Figure 5.2: ......................................................................................................................................... 52 Conceptual diagram of optical multiplexing of four OOK inputs into a 16-QAM. Figure 5.3: ......................................................................................................................................... 52 Experimental setup. CW: continuous wave: MZM: Mach-Zehnder modulator: BPF: band pass filter: PPLN: periodically poled lithium niobate: TDL: tunable delay line. Figure 5.4: ......................................................................................................................................... 53 (a-b) Back-to-back optical eye and constellation diagrams of two 20 Gb/s OOK inputs: (c) optical eye and constellation diagrams of a multiplexed 4-P AM signal: ( d) optical constellation diagram of a multiplexed QPSK signal with EVM of 18.5. x Figure 5.5: ......................................................................................................................................... 54 (a) Optical spectrum at the output of PPLNl with two incoming OOK signals and a CW pump to remove DC from the constellation; (b) Optical spectrum at the output of PPLN2. Figure 5.6: ......................................................................................................................................... 55 (a) Optical spectrum at the output of PPLNl with four incoming OOK signals; (b) Optical spectrum at the output of PPLN2. Figure 5. 7: ............................................................................................................................. ............ 56 (a) Optical spectrum at the output of PPLNl with four incoming OOK signals and a continuous pump as a DC offset remover; (b) Optical spectrum at the output of PPLN2. Figure 5.8: ......................................................................................................................................... 56 Optical constellation diagram of 40 Gbit/s square and star 16-QAM signals generated by multiplexing of four 10 Gbit/s OOKs. Figure 5.9: ............................................................................................................................. ............ 57 Optical constellation diagram of 80 Gbit/s square and star 16-QAM signals generated by multiplexing of four 20 Gbit/s OOKs. Figure 5.10: ....................................................................................................................................... 57 Optical constellation diagram of a 30 Gbit/s diagonal and rectangular 8-PAM signals generated by multiplexing oftlnee 10 Gbit/s OOKs. Figure 6.1: ......................................................................................................................................... 62 Principles of a conventional method to process a QAM signal using an electrical coherent receiver. Figure 6.2: ......................................................................................................................................... 64 Principle of (a) phase insensitive and (b) phase sensitive amplifiers. Figure 6.3: ......................................................................................................................................... 65 Principles for demultiplexing of a QPSK into BPSK sub-channels based on phase sensitive amplification. xi Figure 6.4: ......................................................................................................................................... 67 (a-d) Constellation diagrams of demultiplexed BPSK sub-channels for different gain factors. Figure 6.5: ......................................................................................................................................... 68 BER vs. OSNR of the original QPSK signal and two demultiplexed channels with gain factor of 9.5 dB; Inset: Constellation diagram of the original QPSK signal. Figure 6.6: ......................................................................................................................................... 69 (a) conceptual diagram of 16-QAM generation; (b) Principle of all optical logic/arithmetic gates implementation for 16-QAM sub-<::hannels. Figure 6. 7: ............................................................................................................................. ............ 71 Experimental setup. CW: continuous wave; PC: polarization controller; BPF: bandpass filter; TDL: tunable delay line; HNLF: high nonlinear fiber. Figure 6.8: ......................................................................................................................................... 72 (a) Optical eye and constellation diagrams of 10-Gbaud back-to-back 16-QAM. (b) constellation diagram of 10-Gbaud 16-QAM at HNLFl output. constellation diagram of phase conjugated copy of 10-Gbaud 16-QAM at HNLFl output. Figure 6.9: ......................................................................................................................................... 73 Optical spectrum. (a) at the output ofHNLFl; (b) at the output ofHNLF2. Figure 6.10: ....................................................................................................................................... 74 BER performance versus optical received power. Figure 6.11: ....................................................................................................................................... 74 Error Vector Magnitude (EVM) versus optical received power. Figure 6.12: ....................................................................................................................................... 75 Eye diagram of a 10-Gbaud 3-level amplitude signal. Figure 7.1: ......................................................................................................................................... 78 (a) an optical pulse degraded by amplitude noise in an optical transmission line. (b) Constellation diagrams of an optical signal at different points in an optical transmission line. xii Figure 7.2: ......................................................................................................................................... 79 Experimental setup. FSDL: free space delay line; PC: polarization controller; DD: direct detection; DPSK DM: demodulation using delayed interferometer; DSNR: optical SNR adjustment by adding ASE [83]. Figure 7.3: ......................................................................................................................................... 80 Power transfer function of FP A-based amplitude regenerator ( LP: signal launched average power; RP: received average power) [83]. Figure 7.4: ......................................................................................................................................... 81 ( a-c) Intensity eye diagrams before and after amplitude regeneration; (b-d) demodulated eye diagrams before and after regeneration [83]. Figure 7.5: ......................................................................................................................................... 82 Optical eye diagrams: (a) input signal with 7% amplitude fluctuations; (b) regenerated after a conventional NOLM; (c) demodulated RZ-DPSK signal after a conventional NOLM. (d) Regenerated by the modified NOLM (e) demodulated RZ-DPSK signal regenerated by the modified NOLM [84]. Figure 7.6: ......................................................................................................................................... 83 Power transfer function of NOLM-based amplitude regenerator; Output power and phase shift versus NOLM input power [84]. Figure 7.7: ......................................................................................................................................... 84 Experimental setup for all-optical regeneration and regenerative wavelength conversion of 40 Gb/s DPSK data (LD: laser diode; MZM: Mach- Zehnder Modulator; BPG: bit pattern generator; PC: polarization controller; VOA: variable optical attenuator; EDF A: erbium-doped fiber amplifier; OF: optical filter; SOA: semiconductor optical amplifier; DI: delay interferometer) [138]. Figure 7.8: ......................................................................................................................................... 85 Typical spectrum at SOA output. FWM conversion efficiency is about -13 dB [138]. xiii Figure 7.9: ......................................................................................................................................... 85 Input and output eye diagrams for four-wave-mixing (FWM) and pass-through (PT) signals when input Q-factor is set to 5 [138]. Figure 7.10: ....................................................................................................................................... 86 BER versus threshold margin for four-wave-mixing (FWM) and pass-through (PT) signals when input Q-factor is set to 5 [138]. Figure 7.11: ....................................................................................................................................... 86 Input and output eye diagrams for four-wave-mixing (FWM) and pass-through (PT) signals when input Q-factor is set to 4 [138]. Figure 7.12: ....................................................................................................................................... 87 BER versus threshold margin for four-wave-mixing (FWM) and pass-through (PT) signals when input Q-factor is set to 4 [138]. Figure 7.13: ....................................................................................................................................... 88 Diagram of polarization-based phase-preserving multilevel amplitude regeneration. Figure 7.14: ....................................................................................................................................... 90 Concept of multilevel amplitude regeneration: the original signal is phase-modulated based on the self-phase modulation periodic effect and is added coherently to the original signal. Figure 7.15: ....................................................................................................................................... 92 Simulation setup. CW: continuous wave: PC: polarization controller: l/Q Mod: In-phase/quadrature modulator: AM: amplitude modulator: PBS: polarization beam splitter: HNLF: highly-nonlinear fiber: PBC: polarization beam combiner: BPF: band pass filter: EDF A: erbium doped fiber amplifier: VOA: variable optical attenuator. xiv Figure 7.16: ....................................................................................................................................... 93 Back-to-back constellation diagrams with OSNR 20 dB: (a) a noisy star-8QAM signal: (b) regenerated signal after one-stage amplitude regenerator: ( c) regenerated signal after two-stage amplitude regenerator. Back-to-back constellation diagrams with OSNR 22 dB: (d) a noisy star- 8QAM signal: (e) regenerated signal after one-stage amplitude regenerator: (f) regenerated signal after two-stage amplitude regenerator. Back-to-back constellation diagrams with OSNR 24 dB: (g) a noisy star-8QAM signal: (h) regenerated signal after one-stage amplitude regenerator: (i) regenerated signal after two-stage amplitude regenerator. Figure 7.17: ....................................................................................................................................... 94 Regeneration factor vs. OSNR for both amplitude levels of a regenerated star-8QAM signal after a two-stage amplitude regenerator in a back-to-back configuration. Figure 7.18: ....................................................................................................................................... 95 Power transfer function of the one-stage polarization-based amplitude regenerator. Figure 7.19: ....................................................................................................................................... 96 Back-to-back constellation diagrams with OSNR 22 dB: (a) a noisy star-16QAM signal: (b) regenerated signal after one-stage amplitude regenerator: ( c) regenerated signal after two-stage amplitude regenerator. Back-to-back constellation diagrams with OSNR 24 dB: (d) a noisy star- 16QAM signal: (e) regenerated signal after one-stage amplitude regenerator: (f) regenerated signal after two-stage amplitude regenerator. Figure 7.20: ....................................................................................................................................... 97 Transmitted constellation diagrams of star-8QAM after 150 km transmission line with/without a pre-regeneration module at the input of a nonlinear transmission line. Figure 7.21: ....................................................................................................................................... 98 Transmitted constellation diagrams of star-l 6QAM after 150 km transmission line with/without a pre-regeneration module at the input of a nonlinear transmission line. xv Figure 7.22: ....................................................................................................................................... 99 Phase noise factor vs. input signal OSNR for both amplitude levels of a transmitted star-8QAM with/without pre-regeneration module at the input of a nonlinear transmission line. Figure 7.23: ....................................................................................................................................... 100 Polarizer angle vs. attenuation value embedded in a custom polarization beam combiner (PBC) at BER !Oe-4. Figure 7.24: ....................................................................................................................................... 101 Back-to-back constellation diagrams with OSNR 22 dB: (a) a noisy square-16QAM signal: (b) regenerated signal after two-stage amplitude regenerator. Figure 7.25: ....................................................................................................................................... 103 (a) Optical duobinary detection scheme: (b) Conventional electrical duobinary detection scheme using a DLI. Figure 7.26: ....................................................................................................................................... 105 (a) Standard and optimized Optical eye diagrams of 40Gbit/s NRZ-OOK at !Okm: (b) Optical eye diagrams of an optimized NRZ-OOK to 3-level intensity (duo binary) format conversion for various bit rates at 1555.3 nm. Figure 7.27: ....................................................................................................................................... 106 Dispersion vs. distance at various wavelengths in a 20 Gbit/s system. Figure 7.28: ....................................................................................................................................... 106 Maximum and minimum possible lengths vs. bit rate at which full NRZ-OOK to 3-level intensity modulation format conversion occurs: (insets) Optical eye diagrams of 40 Gbit/s optical duobinary signals at 8km. llkm and 7km. Figure 7.29: ....................................................................................................................................... 107 Electrical NRZ amplitude (volt) vs. maximum reach for NRZ-OOK to duobinary format conversion for various bias voltages in 40 GB/s transmission system with v x = 4. xvi Figure 7.30: ....................................................................................................................................... 109 Diagram of NOLM-free polarization-based phase-preserving multilevel amplitude regeneration. Figure 7.31: ....................................................................................................................................... 112 Simulation setup. CW: continuous wave: PC: polarization controller: l/Q Mod: ill-phase/quadrature modulator: AM: amplitude modulator: PBS: polarization beam splitter: HNLF: highly-nonlinear fiber: PBC: polarization beam combiner: BPF: band pass filter: EDF A: erbium doped fiber amplifier: PS: phase shifter: VOA: variable optical attenuator. Figure 7.32: ....................................................................................................................................... 113 Back-to-back constellation diagrams with OSNR 20 dB: (a) a noisy star-8QAM signal: (b) regenerated signal after one-stage amplitude regenerator: ( c) regenerated signal after two-stage amplitude regenerator. Back-to-back constellation diagrams with OSNR 22 dB: (d) a noisy star- 8QAM signal: (e) regenerated signal after one-stage amplitude regenerator: (f) regenerated signal after two-stage amplitude regenerator. Back-to-back constellation diagrams with OSNR 24 dB: (g) a noisy star-8QAM signal: (h) regenerated signal after one-stage amplitude regenerator: (i) regenerated signal after two-stage amplitude regenerator. Figure 7.33: ....................................................................................................................................... 114 Regeneration factor vs. OSNR for both amplitude levels of a regenerated star-8QAM signal after a two-stage amplitude regenerator in a back-to-back configuration. Figure 7.34: ....................................................................................................................................... 115 Transmitted constellation diagrams of star-8QAM after 150 km transmission line with/without a pre-regeneration module at the input of a nonlinear transmission line. xvii List of Tables Table6.1: ............................................................................................................................. 76 All optical logic gate and arithmetic operation implementations between two sub-channels carrying by a 16-QAM signal. Table 7.1: ............................................................................................................................ 108 All optical logic gate and arithmetic operation implementations between symbols carrying by a 16-QAM signal. xviii 33%RZ 50%RZ amplified spontaneous emission amplitude/phase modulated analog-to-digital converter arbitrary wave form generator balanced receiver band pass filter binary phase-shift keying bit error rate bit pattern generator carrier suppressed RZ chromatic dispersion Continuous wav pump cross phase modulation delay interferometer delay line interferometer demultiplexing List of Acronyms dense wavelength division multiplexed difference frequency generation RZ33 RZ50 ASE PAM ADC AWG BR BPF BPSK BER BPG CSRZ CD CW pump XPM DI DLI DEMUX DWDM DFG xix differential group delay differential phase-shift keying differential quadrature phase shift keying effective number of bits electro-absorption modulator erbium-doped fiber amplifiers error vector magnitude fast Fourier transforms fiber parametric amplifiers for back-to-back forward error correction four-wave mixing four-wave-mixing highly nonlinear fibers In-phase/ quadrature intensity modulated format signals with direct detection intermediate frequency inter-symbol interference laser diode local oscillator Mach-Zehnder modulator DGD DPSK DQPSK ENoB EAM EDFA EVM FFT FPAs B2B FEC FWM FWM HNLF I/Q IM/DD IF ISI LD LO MZM xx nonlinear optical loop mirror non-Return-to-Zero On-Off Keying on-off-keying optical filter optical filter optical parametric amplification optical signal-to-noise-ratio op ti cal/ e 1 ectroni c/ op ti cal pass-through periodically poled lithium niobate waveguide phase modulator phase sensitive amplifiers phase shift keying phase-locked loop polarization beam combiner polarization beam splitter polarization controller polarization division multiplexing polarization-mode dispersion pseudo-random bit sequences pulse pattern generator NOLM NRZ-OOK OOK OF OF OPA OSNR O/E/O PT PPLN waveguide PM PS As PSK PLL PBC PBS PC PDM PMD PRBS PPG xxi quadrature amplitude modulation quasi-phase matching return-to-zero on-off keying second harmonic generation self phase modulation semiconductor optical amplifier semiconductor optical amplifier signal-to-noise-ratio single mode fiber spectral efficiency standard deviation star 8-quadrature-amplitude modulation sum frequency generation tunable delay line variable optical attenuator wavelength division multiplexing zero-dispersion wavelength QAM QPM RZ-OOK SHG SPM SOA SOA SNR SMF SE STD star-SQ AM SFG TDL VOA WDM ZDW xx ii Abstract In the next generation of optical networks, high speed data rates of lOOGb/s per channel or more will be required. In order to include more channels in a single fiber in wavelength division multiplexing (WDM) systems, the channel spacing must be decreased from 200GHz to 50GHz or even smaller. These dense high bit rate WDM systems suffer more severely from the effect of linear and nonlinear degradation in fiber transmission. Phase-modulated format-based WDM systems that have more tolerance for degrading effects, particularly nonlinearity-based degradation have been explored. With growing demand for transmission capacity along with the desire to lower the cost per information bit, spectrally-efficient multilevel modulation format signals such as quadrature amplitude modulation (QAM) have become popular. Higher-order data modulation format signals are quite important to optical communications due to their high spectral efficiency, low electrical baud rate and increased tolerance to fiber-based impairments. Specifically, there is interest in generating quadrature amplitude-modulation (QAM) signals, and researchers have demonstrated up to 1024-QAM. A laudable goal is the generation of high-order QAM in a tunable fashion such that variable bit rates can be accommodated using optical methods. In this dissertation, we demonstrate two different techniques for generation of QAM signals. The first method is a polarization-based technique for high speed, tunable QAM signal generation that implements the amplitude control outside the integrated device using a polarizer. The second method is a nonlinearity-based technique that multiplexes initial lower-level modulation format signals at different frequencies into a QAM signal. We also develop fully optical modules to apply various signal processing operations to QAM xxiii signals, e,g, demultplexing and information extraction as teclmiques to avoid coherent receivers in the middle of optical networks, The optical demultiplexing module is based on the phase-sensitive amplification concept, and can demultiplex a QPSK signal into two BPSK sub-channels, We also develop a fully optical module that provides logic/arithmetic relations between symbols carried by a QAM signaL A disadvantage of higher-order QAM signals is higher sensitivity to noise accumulation, especially in long-haul transmission systems, Amplitude noise not only reduces signal quality but may also be converted into nonlinear phase noise in a transmission line due to the Gordon-Mollenauer effect Cross-phase modulation in wavelength-division multiplexing (WDM) systems is another cause of amplitude noise to nonlinear phase noise conversion, All-optical regenerators are expected to extend the maximum reach of high-speed transmission systems by eliminating accumulated signal impairments in transmission systems without the need for optical/electronic/optical (O/E/O) conversion, Current phase regeneration schemes are relatively complex and are limited to lower-order modulation formats, Thus, developing all-optical tunable phase-preserving multilevel amplitude regenerator modules with bit-rate transparency is desirable, In this dissertation, we demonstrate all-optical multilevel amplitude regeneration for various types of star-QAM and also square-16QAM signals using a teclmique that we call coherent polarization mixing, We describe two different schemes for the proposed polarization-based multilevel amplitude regeneration technique, The schemes provide a wide range of tunability and scalability, and have a simpler configuration compared to previous methods, xx iv Chapter 1: Introduction Classic optical communication systems refer to communication systems that use carriers at very high frequencies (-193.1 THz) and use optical fibers for information transmission. Optical communication systems are classified as direct detection and coherent detection systems based on the receiver type. Intensity modulated format signals with direct detection (IM/DD) have been mainly used in the early days of optical communications because of their simple transmitter and receiver implementation. However this format has a low spectral efficiency (SE) of 1 bit/s/Hz/polarization [l]. There has been dramatic growth in capacity demand in networks. resulting in a simultaneous dramatic increase in the data speeds of transmitters and receivers. This motivates the need to upgrade the existing backbone of communication networks to operate at higher transmission rates. These optical communication networks widely use Dense Wavelength Division Multiplexed (DWDM) transmission systems to increase the transmission capacity by transmitting different data streams on different wavelengths. It is possible to increase the throughput of such DWDM transmission systems by using wider optical bandwidth per channel so that the data rate per channel (baud rate) can be increased. or by using higher-order modulation formats with higher spectral efficiency (SE) that can be used to transmit more information using the same bandwidth 1 respectively [2-6]. Implementing wider optical bandwidth per channel will decrease the number of DWDM channels due to the spectral bandwidth limitation of optical amplifiers [5]. In response to the need for high spectral efficiency, phase modulated signals such as phase shift keying (PSK) and amplitude and phase-modulated signals such as quadrature amplitude modulation (QAM) were developed. Higher-order modulation format signals solve the problem of having a limited number of channels with wider bandwidth in DWDM systems by decreasing the signal baud rate and carrying several bits per symbol [5]. Note that advanced modulation formats have higher tolerance to chromatic dispersion (CD) and polarization-mode dispersion (PMD) as they can reduce the symbol rate while keeping the same bit rate. Consequently, in the 1980s, coherent optical systems gained popularity. At first, an optical phase-locked loop (PLL) was used to lock the phase of the local oscillator (LO) laser to the phase of the received signal, a technique known as homodyne detection [7-8], which is extremely unstable. Later heterodyne receivers were developed that down-convert the received signal to an intermediate frequency (IF) in the microwave region, and then an electrical PLL is employed to lock the phase of the intermediate frequency signal [9-1 O]. Note that the intermediate frequency should be much higher than the signal bit rate. Due to the large line width of lasers of this era and PLL feedback delay, only simple modulation formats such as binary phase-shift keying (BPSK) and differential phase-shift keying (DPSK) [11-12] could be deployed. After developing erbium-doped fiber amplifiers (EDFA) and wavelength-division multiplexing (WDM) techniques in the 1990s, direct detection became popular again for more than twenty years [13]. Coherent systems had been completely ignored until the development of high- 2 speed analog-to-digital converters (ADCs) in 2002 [14-16] which brought back coherent optical systems. Eventually research on the topic of advanced phase and amplitude modulated signals using newly-developed coherent detection technology became popular because it provides the opportunity to retrieve the amplitude and phase information and use offline signal processing to compensate for degradations such as chromatic dispersion (CD). polarization-mode dispersion (PMD) and nonlinearities such as the self phase modulation (SPM) effect [17-22]. Figure 1.1 shows the three optical modules in optical communication systems. The optical QAM generation (multiplexer) module multiplexes several lower-level modulation formats into a higher-order modulated signal. The optical QAM demultiplexing module extracts channel information carried by a higher-order QAM signal. This information may be in the form of logic or arithmetic relations between symbols carried by a QAM signal. or it may be all sub-channels demultiplexed. State-of the-art transceivers are capable of supporting multilevel modulation formats with 100- Gbaud/s Ethernet baud rate per channel. Although these ultra-high-performance transceivers have a large capacity. a laudable goal is to develop transceivers whose capacity can be tailored and shared among many different users by demand. This shows the crucial role of optical multiplexer modules and optical demultiplexer modules in future optical networks. In optical networks the signals must pass a variable number of optical nodes on their way to the destination. As a result. signal degradations can arise during transmission. The degradation is mainly due to noise added by EDFAs as well as noise added in the switching nodes. All-optical amplitude regenerator and phase regenerator modules are a 3 solution to avoid the accumulation of noise, crosstalk and non-linear distortions and to ensure a good signal quality for transmission over any path in all-optical network. A big disadvantage of multilevel modulation format signals such as the QAM format is their high sensitivity to amplitude and phase noise. In particular amplitude noise decreases the distance between different amplitude levels. Amplitude noise also converts into nonlinear phase noise during transmission in optical fiber due to the Gordon- Mollenauer effect. Low bit-rate signal 1 Low bit-rate signal 2 -> Low bit-rate signal 3 Optlcal Genlr11or llocUe ~~tor ',. - ~ ~ Transmission Line ,,7 High bit-rate Multiplexed Signal At a single frequency Node 1 Low bit-rate signal 1 Low bit-rate ,,---~•signal 2 \ Fig. 1.1. Three main fully optical modules in optical communication systems. This effect is one of the major limiting factors for QAM signal transmission. This shows the necessity of removing amplitude noise from QAM signals to maintain signal quality in long haul transmission lines. Note that beside achieving high spectral efficiency, developing fully optical signal processing modules such as a QAM generation (multiplexer) modules, QAM demultiplexer modules and fully optical regenerators may play a key role in future optical networks. Fully optical signal processing modules can avoid the need for optical/electronic/optical (O/E/O) conversion and take advantage of large bandwidth and wide range of tunability in the optical domain [23]. 4 In this dissertation, we demonstrate two different techniques for optical generation of QAM signals. We also develop fully optical modules to apply various signal processing operations to QAM signals, e.g. demultplexing and information extraction as techniques to avoid coherent receivers in the middle of optical networks. We also describe a novel technique to develop a fully optical amplitude regenerator based on a method that we call coherent polarization mixing. It is important that optical signal processing modules operate on advanced modulation formats. In this manner we show that our proposed amplitude regenerator operates on multilevel amplitude modulated signals. We also show that our regenerator has a phase-preserving characteristic with wide range of tunability, which makes it a perfect candidate to perform regeneration on star-8QAM, star-16QAM and square-l 6QAM signals. The subsequent chapters of this dissertation present detailed descriptions of these systems. 5 6 Chapter 2: Basic and Advanced Optical Modulation Format Generation and Detection This chapter provides an overview of different optical modulation formats. It provides an overview of simple modulation formats such as on-off-keying (OOK) and also more complex multilevel modulation formats such as phase shift keying (PSK) and quadrature amplitude modulation (QAM). In this chapter we review the details behind coherent detection concepts as well. We also review electrically-based methods to generate different modulation formats. 2.1 Why Advanced Modulation Formats? The best modulation format candidate depends on various system parameters, including number of users, link length, and system bit rate, cost and intended spectral efficiency (SE). As an example, an advantage of the differential-phase-shift keying (DPSK) modulation format compared to OOK signals in WDM systems is the ability to reduce cross phase modulation (XPM) penalties and its significant improvement in receiver sensitivity when employing balanced receivers [24]. 7 Furthermore, implementation of DPSK along with direct detection is straightforward at 10 and 40 Gb/s bit rates [25-27], The reason behind intense research on advanced modulation formats is because at a fixed optical amplification bandwidth, increasing the transmission capacity requires increasing the spectral efficiency (SE), defined as the net per-channel bit rate divided by the WDM channel spacing !if [28] SE=_!!__, 4f (2-1) Previously, binary intensity or binary phase modulation was sufficient , and the required increase in SE was achieved by advances in high-speed electronics (higher R) as well as in laser and optical filter frequency stability (reduced !if) [28], In the early 2000s, the speed of opto-electronic modulation and detection approached the bandwidth of stable optical filters, therefore higher-order modulation formats were the solution to increase bit rate at fixed signal bandwidth [28], The first popular advanced modulation format was differential quadrature phase shift keying (DQPSK), which allowed point-to-point SE of up to L6 b/s/Hz [29], This SE is measured in a WDM system with 40 Gb/s baud rate and 50-GHz channel spacing [30], The spectral efficiency can be increased to 3,2 b/s/Hz by developing polarization division multiplexing (PDM) in point-to-point applications using direct detection [29], and to 2 b/s/Hz in an optically-routed environment using coherent detection [31], Further compression of the signal spectrum requires higher-level (M-array) modulation formats along with coherent detection, with an SE limited to 2log~ using PDM [28], In practice even by having optically-routed networking with multiple state-of- the-art ROADMs, 50% of this value may be obtained [28], Note that higher SE imposes more signal-to-noise-ratio (SNR) requirements on the optical network [28], Note that the 8 net impact of higher SNR requirements on the achievable transmission reach also depends on fiber type, amplification scheme, and impact of fiber nonlinearities [33-34]. N 10 l-----;---;--~---=-----:::::..__-1 >-~ u V) c: ......_ -~ t:.. u IE 6 Q) :p - ro It! N .b"C u ro v o.. O l/) Q.. 1.... ~ 1 (a) 'N 10 >-~ u IJ) c: ......_ -~ t:.. u IE 6 Q) :,i:i - ro CO N .t:; ·;:: u ro Q) - o..O l/) Q.. 1.... ~ 1 (b) 0 0 8.8 dB 5 10 15 20 25 Required SNR per bit [dB] Constellation 256 _ QAM shaping ?~e==------=:7 64 7 _~Q7,AM ~ " 1 n~ulatlou l6-Q_ AM I 4-QAM _i ···-····- .. L... ·····-··· .l -··· 5 10 15 20 25 Required SNR per bit [dB] Fig. 2.1. Single-polarization spectral efficiency versus the received SNR per bit. The Shannon limit for a linear, additive white Gaussian noise channel is shown together with the theoretical performance of various square QAM formats (blue circles), assuming Gray-coded symbol mapping and state-of-the-art 7% overhead hard-decision FEC. Also shown in (a) are representative experimental results (red squares); numbers indicate QAM constellation sizes. In (b), the Shannon limits for various square QAM constellations are shown, and the effects of constellation shaping, coded modulation, and signal over-filtering are indicated [65]. 9 The trade-off between spectral efficiency and system reach depends predominantly on the underlying modulation format and forward error correction (FEC), and determines the maximum WDM capacity that can be transmitted over a given distance within a practical optical amplification bandwidth [65], Figure, 2,1 (a) shows this trade-off in the linear regime, showing the potential single-polarization spectral efficiency as a function of the received SNR per bit [46-47], The Shannon limit for a linear, additive white Gaussian noise channel [ 48] is shown together with the theoretical performance of different types of higher-order square QAM constellations (blue circles), Note that in this figure the Gray-coded symbol mapping is applied and 7% overhead hard-decision forward-error correction (FEC) capable of correcting an input BER of 3,8 .1Q- 3 to values below 10- 15 is considered [ 49]. Experimental points are shown by red squares. Performance closer to the Shannon limit will require more advanced coding [50-51] and/or implementing non square QAM constellation shaping [52-53]. The impact of advanced coding on linear system performance is quantified in Fig. 2.1 (b). In this figure the blue circles show the performance of the 7% hard-decision FEC underlying Fig. 2.1 (a) and the Shannon bounds for the square QAM formats with ideal soft-decision FEC are shown as blue curves [65]. The asymptotic gap between square QAM performance and the modulation unconstrained Shannon limit (black curve) is 1.53 dB and can be reduced by proper constellation shaping, especially for large constellation sizes [52-53]. Note that the Shannon limit is relatively steep in the low spectral-efficiency regime but asymptotically flattens out to a slope of 1 b/s/Hz for every 3-dB of higher SNR per bit at high spectral efficiencies [65]. As can be seen in Fig. 2.1 (a), by doubling the system capacity (shifting from QPSK to 16-QAM), a 3.7-dB 10 higher SNR per bit, or 6,7-dB higher optical SNR (OSNR) is required [46] under the same symbol rate condition, plus a 1-dB higher expected implementation penalty [65], To get approximately the same QPSK transmission reach for 16-QAM, this SNR gap may be closed by applying techniques such as: (i) applying stronger soft-decision FEC or coded modulation [50], [51], [53] to reduce the receiver's SNR requirements and shift both theoretical and experimental points in Fig, 2,L (a) closer to their Shannon bounds [65]: (ii) using lower-loss fiber or potentially higher-order distributed Raman amplification [54] to increase the OSNR delivered to the receiver: (iii) applying higher optical signal launch powers by using lower-nonlinearity fiber or more powerful nonlinear distortion compensating DSP [46], [55], As can be seen in Fig, 2,1 (a), by doubling the system capacity (shifting from 16- QAM to 256-QAM), 8,8-dB higher SNR per bit is required under the same symbol rate condition, which is impossible to accommodate without reducing system reach [65], The trade-off between spectral efficiency and system reach, including noise, fiber nonlinearities, as well as current technological shortfalls, is discussed in Fig, 2,2, This figure shows the PDM spectral efficiency versus transmission distance, Note that the blue circles show the experimental results [ 46], Both curves are straight lines on a logarithmic scale for the transmission distance L, since the delivered SNR is inversely proportional to L , and the spectral efficiency is given by [65] (2-2) in the high-SNR regime [56], [57], Note that the ellipse shows a range for commercial systems working over installed legacy fiber with appropriate OSNR margins and the asterisk shows Alcatel-Lucent's commercially deployed 1830 optical transmission 11 platform [65], [60]. As can be seen the experimental records have approached the nonlinear Shannon limit to within a factor of less than two. With an annual 2-dB -e 51--~--:-~--=~--...-++-~--'.._~_,,, •~~~..--1 "'"" u :• (I) a. I I V') 2L--~~~~~~~--'-~--7,,f-~~~~----'~-' 100 1,000 10,000 Transmission distance [km] Fig. 2.2. Trade-off between dual-polarization spectral efficiency and transmission reach, showing the nonlinear Shannon limit of [46] together with experimentally achieved results (circles). The ellipse indicates a range into which commercial systems might fall, and the asterisk represents Alcatel-Lucent's commercially deployed optical transmission platform [60], [65]. traffic growth rate, actual traffic will reach the nonlinear Shannon limit in less than 18 months [65]. In another words, as WDM capacities are no longer scalable, alternative solutions have to be speedily developed [65]. 2.2 Electro-Absorption Modulators (EAMs) Electro-absorption modulators (EAMs) are pin semiconductor structures whose band gap can be modulated by applying an external voltage, resulting in changes to the device's 12 absorption properties [36]. EAMs need low drive voltages. The disadvantage of electro-absorption modulators is the generation of residual chirp. Figure 2.3 (a) shows the typical exponential transmission characteristics of an EAM versus drive voltage. The ratio of maximum-to-minimum modulated light power for EAM typically does not exceed 10 dB [35]. Another drawback for EAMs is their limitation in handling optical power. EAM has a back-to-back insertion loss of almost 10 dB. A solution for reducing the high insertion losses of EAMs is the integration with semiconductor optical amplifiers (SOAs) [37]. Power transmission [dB] 0 -10 -20 Drive voltage (a) Power transmission [%] iQ.~OUt V 2 (t) Voltage difference (t. V) (b) Fig. 2.3. Transmission functions of: (a) EAMs and (b) MZMs [35]. 2.3 Mach-Zehnder Modulators (MZMs) Unlike EAMs, whose function is based on the principle of absorption, Mach-Zehnder modulators (MZMs) operate on the principle of interference that is controlled by modulating the optical phase [35]. The modulator structure is shown in the inset to Fig. 2.3 (b). The initial light is split into two branches at an input coupler. Light in both paths can be phase-modulated which results in a phase delay controlled by the applied voltages 13 v 1 and v 2 • The two optical fields then interfere destructively or constructively in the optical coupler [35]. The optical field transfer function TE(vp v 2 ) of the MZM is [38] T(v v)=-Lm,,l+e1¢(,,)+1"'}=/( 2 )cos \l'v1 -\l'v2 'Jf 1 ¢(,,)+¢(,,)+'!' [((d.( ) d.( )) ] E1'2 2t" 2 2 (2-3) where ¢(v 1 ) and ¢(v 2 ) are optical phases of the two paths of the MZM. Here 'Jf is a constant phase shift in one of the arms. which is called modulator bias. If the relation between phase change and voltage is linear. which is true for most materials used for MZMs. the power transfer function is [35] (2-4) where ¢=av and a is a constant. The MZM power transfer function is shown in Fig. 2.3 (b ). V, is the voltage difference required to change the phase between arms by Jr • Another degree of freedom in choosing v 1 + v 2 can be used to apply phase modulation. which is called signal chirp [39-41]. If we want to ignore the chirp factor. we will have (2-5) and the phase term in Eq. (2-3) vanishes. This driving condition is known as balanced driving. Some modulator structures. e.g .. x-cut lithium niobate. have a single drive. that modulates both arms and balanced driving condition for chirp-free condition achieved by appropriate drive electrode design [35]. Figure 2.4 demonstrates an overview of the different options of driving an MZM to generate different modulation formats [35]. 14 ........ Differential voltage (~ V) Time T = 1/R Fig. 2.4. Overview of different options in driving an MZM (Black circles: MZM quadrature points) [35]. 2.4 On-Off Keying (OOK) As shown in Fig. 2.3, the MZM modulator should be biased at 50% transmission or at one of the quadrature points and is driven from minimum to maximum transmission with a voltage swing of v "' to generate non-Return-to-Zero On-Off Keying (NRZ-OOK). Note that black circles in Fig. 2.4 define quadrature points. Figure 2.5 shows the optical spectrum and the optical intensity eye diagrams of an NRZ signal. The optical spectrum includes a continuous portion, which is the result of the shape of the individual NRZ data pulses, a strong tone at the carrier frequency, and weaker tones at multiples of data rate (R) intervals. 15 ~ F;~~mm IX , XI 0 ···········:··········· ·· ···-·····:··········· Q) c.. en ~ :g_ 0 R ~ Frequency Fig. 2.5. Optical spectrum and optical intensity eye diagrams of NRZ-OOK signal [35]. Another basic optical modulation format is return-to-zero on-off keying (RZ-OOK). Unlike NRZ, in all RZ modulation formats, the adjacent marks are separated by periods in which the magnitude returns to the low level. Optical RZ formats can have different duty cycles. Three of the most typical RZ formats are 50% RZ (RZ50), 33% RZ (RZ33) and carrier suppressed RZ (CSRZ). As shown in Fig. 2.4, RZ-OOK is generated by carving NRZ signal pulses by using another MZM modulator which is called pulse carver. Pulse carver modulators are MZMs driven by sinusoidal clock signals. For different RZ modulation formats clock recovery is easier, inter-symbol interference (ISI) effects are less visible and higher sensitivities can be achieved [140-143]. However, as RZ formats have wider spectra, they fail in achieving a high spectral efficiency. A further consequence is that RZ formats will suffer from greater amounts of optical filtering penalties since filter concatenation effects will be more severe [144]. 2.5 Differential Phase Shift Keying (DPSK) Differential Phase Shift Keying (DPSK) encodes data on the binary phase change between adjacent bits. Binary bit 1 is encoded into a ;rphase shift and binary bit 0 is 16 encoded by having no phase change [35]. DPSK can also be in NRZ and RZ format. The advantage of DPSK over OOK is a 3-dB receiver sensitivity improvement [42-43]. (a) (b) Fig. 2.6. Constellation diagrams of (a): OOK; (b): Differential phase shift keying (DPSK) [35]. As can be seen in Fig. 2.6 (b ), for the DPSK modulation format, the symbol spacing is increased by .Ji compared to OOK modulation format considering fixed average optical power [44]. This means that for the same BER, DPSK modulation format can tolerate .Ji times more standard deviation of optical noise than OOK modulation format. This is equal to a 3-dB reduction in OSNR [35]. An optical DPSK transmitter is shown in Fig. 2.7. To avoid error propagation, the data sequence is differentially encoded at the transmitter. The preceding function is represented by the two bit patterns in Fig. 2.7. At the second stage the continuous optical power is phase modulated between 0 and n by the preceded binary data sequence. Phase modulation function can either be applied by a phase modulator (PM) or a mach-zehnder modulator MZM [35]. 17 Intensity I IPM Laser f'C5jjj MZM PM or MZMt-----------c Phase I c:J n IPM I r=i D tMZM lol1o~o Data Precoder Fig. 2.7. RZ-DPSK trans mitter based on two different approaches: (1) Phase modulator , (2) Mach-Zehnder modulator (MZM). Intensity and phase waveforms of PM-based method and MZM-based method are different [35]. In PM-based method, phase is modulated along the unit circle, which means there is no intensity change. As shown in Fig. 2.7, in PM-based method, optical phase completely follows the electrical drive signal, and phase transition speed is limited by the bandwidth of electrical amplifier and phase modulator. As shown in Fig. 2.6 (b), in MZM-based method, the MZM is symmetrically driven around zero transmission, and the phase modulates along the real axis through the origin of the constellation plane which results in generating the exact n phase [35]. Fig. 2.8 (a) and Fig. 2.8 (b) show optical spectrum and intensity eye diagrams of NRZ- DPSK and 33% RZ-DPSK. The absence of a binary bit 0 can be seen in the eye diagrams as a main representation of DPSK signals. The intensity dips between two adjacent bits in the NRZ-DPSK eye diagram is another demonstration of the residual intensity modulation of MZM-based method. 18 E = ;:; a..> = <J? Frequency (a) E = ~ ().) = en --· ---- --} ··--- --- --· -- F requency (b) Fig. 2.8. (a):Optical spectrum and optical intensity eye diagrams of NRZ-DPSK signal; (b): Optical spectrum and optical intensity eye diagrams of 33% RZ-DPSK [35]. 2.6 Differential Quadrature Phase Shift Keying (DQPSK) Modulation formats that can carry more than one bit per symbol are known as multilevel modulation format signals. A common multilevel modulation format is differential Fig. 2. 9. Constellation diagram of the optical field of differential quadrature phase shift keying (DQPSK). quadrature phase shift keying (DQPSK). As can be seen in Fig. 2.9 the constellation diagram for DQPSK consists of four phase shifts {o 70 70 - 70 }. For this multilevel '2' ' 2 modulation format, the symbol rate (baud rate) is half the system bit rate. A DQPSK transmitter is traditionally built by using two parallel MZMs that operate as phase modulators. Figure 2.10 shows a DQPSK transmitter in which light splits into two paths with the same power. Each path contains an MZM that operates as a DPSK transmitter. By having a Jr phase shifter in one path and combining the two paths coherently (which 2 comes from the fact that both paths have exactly the same length), the DQPSK multilevel 19 modulation format can be generated [35 ]. This modulator is known as in-phase/quadrature (I/Q) modulator. The conste llation diaNams of the upper and lower paths and the output are also shown in Fig. 2.10. -0 <I> -0 (I) ,..----.. 8 ro Laser <I> -0 ..... a.. ,lm{E} 1 -r---\ ·-f : t-Re{E} \ ! / J.lm{E} 1--- "-- , '--·' ~ ' ...... ~ '-' ; ·f* ; /f,· >---+ -~ -; -- }-Re{E} 7t/2 ·~!TI{E} ':.{; :Y /V! I ......... -r-· . \ ·-{~--- ·---;Re{E} ~-··"'~ ' Fig. 2.10. DQPSK transmitter (In-phase/quadrature (I/Q) modulator): Two MZMs operate as phase modulators and their outputs combine after applying 90°phase shi ft on one arm. This modulator is known as an I/Q modulator [35]. As generating multilevel electrical drive signals is a complicated task, dependency of this transmitter setup on binary electronic drive signals is considered a big advantage. Like the previous modulation formats, RZ-DQPSK can also be generated by implementing a pulse carver after DQPSK transmitter. The optical spectrum and intensity eye diagram for NRZ-DQPSK and 50% RZ-DQPSK are shown in Fig. 2.11 (a) and Fig. 2.1 1 (b). 20 Frequency (a) E 2 0 Q) CL en ~ :g_ 0 --.. - - .... -~ ..... --... - .. F requency (b) Fig. 2.11. (a): Optical spectrum and optical intensity eye diagrams of NRZ-DQPSK signal; (b): Optical spectrum and optical intensity eye diagrams of 50% RZ-DQPSK [35]. Note that the DQPSK optical spectrum is identical to that of DPSK and the only difference is that the DQPSK spectrum is compressed in the frequency domain by a factor of two. The compressed spectrum is the goal in today's optical communications as it increases spectral efficiencies in WDM systems [35]. The other advantage of having a compressed spectrum is its greater tolerance to chromatic dispersion (CD) [35]. 2.7 16-Quadrature Amplitude Modulation (16-QAM) As it is mentioned modulation formats that carry more than one bit per symbol are known as multilevel modulation format signals. Another popular multilevel modulation format is 16-Quadrature Amplitude Modulation format (16-QAM). As can be seen in Fig. 2.12, the constellation diagram for 16-QAM is more complex compared to DQPSK. In the DQPSK format, information is modulated only in phase, while in the 16-QAM format information is modulated in phase and amplitude simultaneously. Specifically for 16- QAM signal, the constellation consists of twelve different phases and three different amplitude levels. 21 lm{EJ 0 0 0 0 0 0 0 0 0 0 0 0 Re{Ex} 0 0 0 0 Fig. 2.12. Symbol diagrams of the optical field of 16-Quadrature Amplitude Modulation (16-QAM) constellation. For this multilevel modulation format the bit rate is four times larger than the symbol rate (baud rate). Shown in Fig. 2.13 is the most common way to generate a 16-QAM signal based on implementing two parallel in-phase/quadrature (I/Q) modulators. As explained in previous sections, using an I/Q modulator is a common method to generate QPSK signals. In 16-QAM transmitter, two inline QPSK signals generated by two I/Q modulators are combined at the output after applying 6dB attenuation on one of the QPSK signals in order to generate square 16-QAM signal. Again, because generating multilevel electrical drive signals is a complicated task, dependency of this transmitter setup on binary electronic drive signals is considered as a big advantage. One disadvantage for this transmitter is that fabricating such structures with precise amplitude control is complex for high bit rates due to electronic limitations. Due to bit rate 22 ---- ·- -- . .... .... . ........ ---- ---- ---- ATT ~ ·- -- :t! ~--~ 1 ' ..,... I Data3 ..... '"" ...... 1 - - -'9 ATT=6dB Da1:a 4 Fig. 2.13. 16-QAM transmitter: including two parallel (In-phase/quadrature (l/Q)) modulators and a6dB attenuator. limitatioos and fabrication limitations of opto-electrical based 16-QAM transmitters, fully optical methods for generating 16-QAM signals are desiralie. 2.8 Coherent Receiver Hardware and Algorithms The opto-electronic front-end of a intradyne receiver is shown in Fig. 214. First the coming signal is mixed with a local oscillator (LO) laser ) in a polarization-diversity 90-degree optical hybrid. Note that the LO is tuned to within approximately MHz of the coming signal's frequency. The four 90-degree optical hybrid's outputs are detected by a pair of balanced receiver for both polarization states. Then the four outputs that are as~iated with in-phase (x-pol arization state) Ix , quadrature (x -polarization state) Q,, , in-{ilase (y-polarization state) I 'I and quadrature (y-polarization state) QY are sampled and asynchronously digitized at 50 GSamples/s using a commercial 4-channel real-time oscilloscope, having analog-to-digital converters (ADCs) with a 8-bit resolution and with a frequency-dependent effective number of bits (ENoB) between 4 and 5 [28]. 23 ,,,-------------- .... - ---... ---.- ..... " " ' I I I ' I I~ ' - on ' 90-deg I Vl c: Q; I I l( I - ·v; I hybrid ·~ Cl. Vl I I ADC I ~cl 1l I Signal ' PBS 0.. I Vl e I c: Cl. ~ .!2 <lJ 90-deg ::::: c: LO laser ~ ·- - I I - hybrid ~ "l""" - 0 I ' " " ..... _____________ .,,. , ______ _ Polarization-diversity Real-time 90-degree hybrid oscilloscope Fig. 2.14. Setup of the polarization-diversity coherent receiver [28]. As shown in Fig. 2.15, The four ADC outputs are the real and imaginary parts (I and Q) of two complex sample streams, coming from two orthogonal polarization states. In the first block, an algorithm performs ac coupling and corrects for specific (static) optical front-end errors, such as a sampling skew between I and Q signal components or phase errors within the 90-degree hybrid [28]. In the second block, digital anti-aliasing filtering is applied on sample streams. In the third block, a linear filter is implemented in the frequency domain using fast Fourier transforms (FFT) and multi.plication with the quadratic spectral phase characteristic of chromatic dispersion (CD) to compensate for transmission line dispersion. Then the clock frequency is recovered by taking the FFT of the signal power waveform and detecting the (e.g. 14-GHz) tone [28]. By having the 24 Ix Vl c on c ... 0 c 0 Q) ·v; · - -z ~ ... ~ Q_ u ;,;: Q) Q) E Q) ... c.. > ltl ... on VI 0 Vl 0 c :0 u Q) u ·v; Q) ... ly ltl u ... ~ "'O · r:; ~ on c (I) ltl ltl u c ' E 0 · - ' E µ · .;::::; - c c 0 u :.;::::; ... 0 <{ ~ Q) ... LL. u ~ Fig. 2.15. Block diagram of the digital signal processing used within the coherent receiver [28]. recovered clock frequency, the signal is down-sampled to a synchronous 28 GSamples/s (2x oversampling at 1/T = 14Gbaud). Note that the clock phase is still unknown. Finally, a series of complementary detection algorithms perform as well [28]. 25 26 Chapter 3: Cascaded Second-Order Nonlinear Techniques for High Speed Reconfigurable Optical Add-Drop Multiplexing (ROADMs) This chapter provides a brief overview of different nonlinear effects in a device called periodically poled lithium niobate (PPLN) waveguide. The PPLN device is one of the main tools in optical signal processing. The optical techniques used for signal processing in this device includes: sum frequency generation (SFG), second harmonic generation (SHG) and difference frequency generation (DFG). 3.1 Introduction Some optical signal processing functions such as optical add-drop multiplexing for advanced modulation formats can be implemented based on nonlinear effects that occur in PPLN under certain conditions. 27 The following sections focus on the mechanisms of the PPLN device, as well as the processes in this nonlinear device that are used to realize the system functionalities presented in Chapter 5. 3.2 Cascaded Second Harmonic Generation and Difference Frequency Generation (cSHG/DFG) In this chapter, we give an analytical model of PPLN waveguides. Our description here is largely based on the discussion in reference [61]. As can be seen in Fig. 3.1 (a) for cSHG/DFG processes, continuous pump (CW pump) and main signal are launched into the PPLN waveguide to generate a new idler. -------- DFG , ..... , ,.. .. ~# ', " .. , 1' -;,,..._--- - ..... ... , I "~ \ '""" ' I I ' SHG '~ Ws Wp QPM Pump Signal\ · '~, ------------- \- 'd'_ ··za~~ ~ Wp Ws w, WsH QPM (a) (b) Fig. 3.1. Mechanism of cSHG/DFG processes in a PPLN waveguide: (a) The CW pump wavelength set near the QPM wavelength; (b) The signal wavelength set near the QPM wavelength [61]. As optical waves propagate along the PPLN waveguide, the CW pump is the optical wave that creates a SH idler at the doubled frequency, and the main signal interacts with the SH idler simultaneously to generate a new idler based on the difference frequency generation (DFG) process at cv 1 dle~ = 2cvcw -cvsienal. As can be seen in Fig. 3.1 (b) for cascaded SHG/DFG processes, continuous pump (CW pump) and main signal are 28 launched into the PPLN waveguide to generate a new idler. However in this modified mechanism, the main signal wavelength set near the QPM wavelength, As optical waves propagate along the PPLN waveguide, the main signal is the optical wave that creates a second harmonic wave at the doubled frequency and the CW pump interacts with the SH idler simultaneously to generate a new idler at wml& = 2.wdgnaz- wcw based on the difference frequency generation (DFG) process, 3.3. Analytical Model of a Periodically Poled Lithium Niobate Waveguide In this section we review the analytical model of cascaded second order nonlinear interactions in a PPLN waveguide, For simplicity, we assume that the optical waves in the PPLN waveguide can be treated as propagating along z axis in a single waveguide mode, Thus the electric fields of different optical waves are represented as E/x, y,z,t) = ~e/x, y)E/z,t)exp(ik 1 z-iw/)+e,c, (3-1) where the subscripts j = S, P, SH, i refer to the signal, pump, second-harmonic wave, and idler signal, respectively and e,c, denotes complex conjugate terms, Here e 1 (x, y) is the normalized transverse field profile defined by # eJ(x, y)dxdy = 1 (3-2) (-co,co) E/z,t) is the complex amplitude of electric field and k 1 is the propagation constant at frequency m 1 , Here E 1 (z, t) can be normalized as A/z,t)= Aeffn -- 1 E(z,t) 2Cfiu 1 (3-3) 29 where A, 11 is the effective nonlinear interaction area, n j is the refractive index, Jiu is the permeability of free space, and c is the light velocity in vacuum, Note that such normalization represents the optical power as (3-4) where s 0 is the permittivity of free space, Here we assume a slowly varying envelope approximation, Under this assumption the electric field amplitude changes slowly relative to the fast optical carrier frequency, and we derive the following coupled-mode equations where aAP a aAP i /3 a 2 Ap 1 A , A 'A c· A ) --+PIP--+- 2P-2-+-ap p =l(J)pKsHG p SH exp lLlKsHoZ az at2 at 2 aASH aASH i a 2 AsH 1 -- + f31sH --+- f32sH --2-+ - aSHASH az at 2 at 2 = ~l))SHKSHGAPAP exp(-iAfSHGz) + iwSHKDFGAsA; exp(-iLiKDFGz) (3-5) (3-6) (3-7) (3-8) 30 (3-9) (3-10) (3-11) Ak 2Tr nSH np 1 LJ =k -2k --=2Tr[--2--- 1 SHGSH PA A AA SH p (3-12) (3-13) In the above equations, As, AP, AsH and A; are functions of the position z and time t and denote the normalized complex amplitudes of the signal, pump, second-harmonic wave, and converted idler wave, respectively, /JlJ and jJ 21 are the first and second derivatives of the propagation constant k 1 with respect to the angular frequency w , aj is the waveguide propagation loss coefficient and KsHa, KnFa, L1KsHa and L1KnFa refer to the SHG coupling coefficient, DFG coupling coefficient, SHG phase mismatching, and DFG phase mismatching respectively, Also def! = 2 d 3 /Tr is the effective nonlinear coefficient and A is the period of the periodically poled structure in the PPLN waveguide, In order 31 to get analytical solutions, we consider the CW operation (a/at = 0, a 2 1at 2 = 0) under the non-depletion approximation during the cascaded SHG and DFG processes and assuming a lossless waveguide, Thus we obtain IAsl,IAPI >>IA;I As(z)= As(O ),Ap(z)= Ap(O ), (3-14) The boundary conditions of AsH = 0, A,( 0) = 0 and the waveguide length of L are adopted, In addition, the pump wavelength is set at the SHG QPM wavelength so as to satisfy the SHG QPM condition, i,e, LIKsHG = 0, As a result, Eqs, (3-5), (3-7) and the second term in Eq, (3-6) are negligible, Then, we can solve Eq, (3-6) as (3-15) By substituting Eq, (3-15) into Eq, (3-8) we derive (3-16) The corresponding optical power can be obtained from Eq, (3-16), i,e, (3-17) where 32 j( A( L)= L 2 . 2( L1KDFGL) LJ DFG' 2 Slll C iJKDFG 2 2L 2 L 2 --- 2 sinc(L1( DFGL)+ 2 iJK DFG iJK DFG (3-18) Obviously. we can achieve the expression (3-19) with respect to the normalized complex amplitudes according to Eq. (3-16). We also deduce the relationship (3-20) among optical powers from Eq. (3-17). which shows that the power of the converted idler is linearly proportional to the input signal power. 3.4. Cascaded Sum Frequency Generation and Difference Frequency Generation (cSFG/DFG) As can be seen in Fig. 3.2 for cSFG/DFG processes. a continuous pump (CW pump). main signal and a dummy continuous pump are launched into the PPLN waveguide to generate a new idler. The CW pump and main signal set at the same distance from QPM wavelength. As optical waves propagate along the PPLN waveguide. the CW pump and the main signal are the optical waves that are used to build an idler signal at frequency 33 x :z ·.···········1······1 i ii fsFG fdummy /pump ' /signal f;dler ' .1 QPM /\. ¥ ( ) •. ······· x 2 DFG ....... . SFG· f ........ f ............. +·J··· · SFG signal pump DFG: fconverted = fsFG - fdummy Fig. 3.2. Mechanism of cSFG/DFG processes in a PPLN waveguide: The CW pump and main signal set at the same distance from QPM wavelength. msm = mew + msignaland the dummy continuous pump interacts with the generated SFG idler simultaneously to generate a new idler at m;d1er = msFG - mdummy based on the difference frequency generation (DFG) process. Theoretically both cSHG/DFG and cSFG/DFG mechanisms can be used for various optical signal processing. However, it is shown that the same wavelength conversion efficiency can be achieved by employing the cSFG/DFG scheme even though much lower pump power is used for each pump source, as compared with the cSHG/DFG scheme. Note also that the tolerance of temperature for the cSFG/DFG scheme remains the same as that of the cSHG/DFG scheme. On the other hand, in cSFG/DFG configuration, the pump wavelength tuning range is not as critical as that for cSHG/DFG one. Therefore, it is easy to control the pump light. Moreover, the 3-dB signal conversion bandwidth in the cSFG/DFG configuration is broader than that in cSHG/DFG one. It should be mentioned that cSFG/DFG wavelength conversion scheme is very attractive for practical applications [102]. 34 Chapter 4: High Speed Reconfigurable Polarization Based Optical Multiplexing Technique for QAM Signal Generation In this chapter we describe a new polarization-based technique for optical multiplexing of several individual lower-order modulated signals at the same wavelength into one higher- level modulated signal. As a proof of concept, we also experimentally show tunable optical generation of up to 40-Gbaud 16-QAM signals by coupling of two QPSK signals at orthogonal polarization states using a differential group delay (DGD) element. We also show the baud rate tunability of this technique and investigate the tolerance of 16-QAM generation to the offset value of the DG D element. 4.1 Introduction Higher-order data modulation formats have become quite important to the optical communications community due to their high spectral efficiency, low electrical baud rate and increased tolerance to fiber-based impairments. Specifically, there is interest in the generation of quadrature-amplitude-modulation (QAM) and researchers have shown up 35 to 1024-QAM [66]. Typical methods for generating 16-QAM include: (a) electrical 16- QAM generation using I-Q modulator. which can be synthesized by a pulse pattern generator (PPG) or an arbitrary wave form generator (AWG); (b) optical 16-QAM generation by combining two quadrature-phase-shift-keying (QPSK) signals with different amplitude levels in an optical interferometer; and ( c) 16-QAM generation using phase-stabilized fiber interferometer [62-64]. In general. these 16-QAM generation techniques are opto-electrical and tend to be fixed in terms of bit rate and have a narrow range of tunability. Thus. a desirable goal is the generation of higher-order QAM signals in a widely tunable fashion such that variable bit rates can be accommodated. A promising way to achieve these goals is the generation of QAM signals using fully optical techniques or other techniques with more degrees of freedom for tunability. In the optical domain we can take advantage of large bandwidth and wide range of tunability. In this chapter. we demonstrate 10-40-Gbaud baud-rate-tunable optical generation of 16-QAM by coherent addition of two QPSK signals with orthogonal polarization states. Bit error rate (BER) measurements are performed to evaluate the quality of generated 16-QAM at 10. 15. 20. and 40-Gbaud. 4.2 Concept Figure 4.1 shows one of the conventional opto-electrical methods for 16-QAM signal generation. In this scheme one of the QPSK signals is attenuated by 6 dB before coupling with the other QPSK signal. As a result, a square 16-QAM is generated. However, a precise fabrication is required to set this amplitude difference; otherwise it will cause difficulty in generating a square standard 16-QAM. The method also has a small range of 36 tunability, requiring a completely new device to be able to generate a QAM signal with higher baud rate or higher spectral efficiency . Data 1 Data2 Data 4 ·- -· I ' " I '" ' I -· •••• •••• •••• •••• L " ' I ·- -· ATT=6dB Fig. 4.1. Concept of a conventional electro-optical method to generate a 16-QAM signal. To overcome these difficulties, we describe a polarization-based scheme in which two QPSK signals are coupled by a polarization beam combiner (PBC). At the output of the polarization beam combiner, the initial signals are multiplexed at two orthogonal polarization states. As shown in Fig. 4.2, at the final stage, the amplitudes of two QPSK signals at orthogonal polarization states add coherently with each other in a polarizer device. As can be seen, there is no need for a variable attenuator for amplitude control as this can be done by rotating the polarizer angle to adjust the required amplitude difference between two QPSK channels. The electric field related to initial QPSK channels is given by EQPsKi = Re{Aa 1 exp (i UJt - ikz - i CfJi (t)} EQPsK 2 = Re{Ali 2 exp(icut-ikz-i (/)2, (t)} (4-1) where a 1 and a 2 are polarization states of QPSK channels that are orthogonal. These two signals are multiplexed after passing through a PBC. The output of PBC goes through a 37 polarizer to achieve multiplexing of two QPSK channels into a 16-QAM signal. As can be seen in Fig. 4.2, coherent interference between two QPSK signals can be interpreted as coherent vector addition between one point from QPSK 1 and a second point from QPSK 2, resulting in a new point on 16-QAM constellation plane. Note that the polarizer is not only used for multiplexing two QPSK signals, but also adjusts the amplitude ratio between initial QPSKs. An amplitude ratio of two is required between two QPSK copies to generate a 16-Q AM signal. Dala1 CW Laser y polarization .<, • • Eyo Data3 • • Data4 ~ . • ~\ • • E,o QPSK1 Polarization beam combiner Polarizer Squar&16-QAM •••• •••• •••• • ••• •------------------------------------------------------------------- Fig. 4.2. Concept of the polarization-based QAM generation method. The Jones matrix of a polarizer with an axis at <p versus xis given by [48] [ Ex]= [ cos 2 (</J) cos(</J)sin(</J)][Ex o] Er cos(¢) sin(¢) sin 2 (</J) Ero (4-2) where Exo and E yo represents the electric fields at the output of the polarization beam combiner carrying two individual QPSK channels respectively. As can be seen from this 38 equation the desired amplitude ratio to generate a square 16-QAM signal can be achieved by oojusting the polarizer angle such that ¢ = 265° or 63.4°. (4-3) Figure 4.3 depicts a conceptual diagram of 16-QAM generation by the proposed polarization-based technique. As shown in Fig. 4.3 (a), each point of 16-QAM in the I!Q plane can be interpreted as a coherent vector addition of one point from QPSK 1 and a second point from QPSK 2. Note that the vector magnitude of QPSK 2 should be twice (6-dB power difference) that of the QPSK 1 magnitude in order to generate square 16- QAM. Shown in Fig. 4.3 (b) is a simple passive approach to generate 16-QAM using a tunable DGD element as a proof of concept for our proposed polarization-based method. • i• • ,----------------------------~ . . : QPSK 1 i .r . - ~~" : + Square-16-QAM •••• •••• •••• •••• !• T • , :-;r-;+ ' . ' : QPSK2 ' : (a) i ,_ ---------- ------- ---------- _ , • ••• • ••• (b) Fig. 4.3. Concept and principle of (a) vector addition and (b) tunable 16-QAM generation. 39 The input QPSK is projected to the slow and fast axes of the DGD element, which provides two copies at orthogonal polarization states. Initial QPSK polarization state is tuned by a polarization controller (PC) along 45° versus slow axis before entering the DGD element. The relative delay between the two copies of QPSK can be applied by varying the DGD value which makes these two copies uncorrelated. When the relative delay is an integer multiple of the symbol time slot and the vector magnitude of QPSK 2 is twice that of the QPSK 1 magnitude by tuning the polarizer, a coherent vector addition of the two QPSK copies through a polarizer device results in generation of 16-QAM. 4.3 Experimental Setup and Results The experimental setup is shown in Fig 4.4. the QPSK signal is generated by a 10-kHz linewidth continuous-wave (CW) laser at -1550 nm and an In-phase/quadrature (I/Q) modulator with a bandwidth of - 40 GHz, which is driven by two 2 7 -1 pseudo-random bit sequences (PRBS) at 10-40-Gbaud. The generated QPSK polarization is adjusted by a polarization controller (PC) and then is launched into a tunable DGD element with a maximum DGD of 220 ps and a resolution of 0.01 ps. The delay is tuned to 200 ps in order to make QPSKs at orthogonal states of polarization decorrelated by multiple symbols. At the output of the DGD element, these two signals go through a PC and are then sent through a polarizer in order to multiplex these two channels into a 16-QAM, where the launching angle is adjusted by the polarization controller before the polarizer. Due to the -20 GHz bandwidth limitation of analog-to-digital converters (ADC), the generated 16-QAM signal is sent through an electro-absorption modulator (EAM) for 40 demultiplexing. The 40-Gbaud 16-QAM signal is demultiplexed to 10-Gbaud after the EAM, and the demultiplexed 16-QAM is sent to a coherent detection test set for BER measurements. :····················································································t: 1 QPSK Transmitter 1 ! Data2 .i. Data 1 ! I L;~r MZM "':::::=..tl....,.i""""":-.i.:~ L ................................................................................... J. PC Tunable DGD Element -, I 40 Gbaud Only I L--------• Fig. 4.4. Experimental setup. CW: continuous wave; PC: polarization controller; MZM: Mach Zehnder modulator; DGD: differential group delay; EAM: electro-absorption modulator; VOA: variable optical attenuator; BPF: band pass filter; LO: local oscillator; BR: balanced receiver; ADC: analog-to-digital converter. Figure 4.5 shows the optical spectrum of the generated 40-Gbaud 16-QAM before demultiplexing. First we vary the baud rate of the QPSK signal and verify the proposed Fig. 4.5. Experimental Optical spectrum of 40-Gbaud 16-QAM. 41 16-QAM generation polarization-based technique at 10, 15 and 20-Gbaud, without using the demultiplexing stage. The curves in Fig. 4.6 show the BER performance of the generated 16-QAM at three different baud rates of 10, 15 and 20-Gbaud. - a:: w al 2 ~ 3 en 0 -t 4 -10-Gbaud - 15-Gbaud ____.,,...___ 20-Gbaud -42 -34 -26 -18 Received Power (dBm) Fig. 4.6. Experimental results. BER performance of 10, 15, 20-Gbaud 16-QAM signals versus received power. The curves in Fig. 4.7 show the BER performance of the generated 16-QAM at three different baud rates of 10, 20 and 40-Gbaud after demultiplexing to 10-Gbaud respectively. 0:: w al 2 ~ 3 en 0 __J I 4 -42 -34 -26 -18 Received Power (dBm) Fig. 4.7. Experimental results. BER performance of 10,20,40-Gbaud 16-QAM signals after demultiplexing to 10-Gbaud versus received power. 42 The demultiplexing process introduces -2 dB power penalty for 20-Gbaud signal and -4 dB for 40-Gbaud signal. Figure 4.8 (left) shows the constellation and eye diagram of 40- Gbaud QPSK signal after demultiplexing to 10-Gbaud and Fig. 4.8 (right) shows the constellation and eye diagram of generated 16-QAM signal after demultiplexing to 10-Gbaud. We also investigate the effect of DGD offset on the performance of 16-QAM 40--G-b.audl QPSK 40-G-b.aud! 1 6 Q.A:\..I! Co~ [E!ll'B:tio:m&E~~ .after Co:!» [E!ll.atio:ni&:.E ye .aftl!!:r- DE::\.n:::x to 1 1(11..;G-b.aud! DE::\.R::X t o 11()...G.b.aud! ... : J ....... - · • JI!' .. .. .. ...... . ""'- ... • , _... .. • .,.: ... lid- II I Fig. 4.8. Experimental constellations and eye diagrams of 40-Gbaud QPSK and 16- QAM after demultiplexing to 10-Gbaud. DEMUX: demultiplexing. generation. When we set the DGD value to 200 ps and change the bit rate, we observe a certain penalty. As shown in Fig. 4.9 (left), when the DGD value is offset by -13 ps for 10.7-Gbaud 16-QAM, a power penalty of -3dB at 10- 3 BER is observed. Similarly, when the DGD value is offset by -33 ps for 12-Gbaud 16-QAM, a power penalty of -9dB at 10- 3 BER is observed. As can be seen the power penalty can be compensated by fine tuning the DGD value for each baud rate successfully. The insets show the constellation diagram with accurate DGD value required at 12-Gbaud and the constellation diagram 43 - 0:::: w m 2 ~3 C> 0 __. I 4 ---e-- 10-Gbaud --e-- 10.7-Gbaud (DGD fine-tuned) ~ 12-Gbaud (DGD fine-tuned) ----- 10.7-Gbaud (13 ps DGD off) ~ 12-Gbaud 33 s DGD o -42 -34 -26 Received Power (dBm) -18 12-Gbaud w/o DGD Off set .... ... • .. .. .. "'" '*" • - , .. "ll1f' .. .. • 12-Gbaud w/ 33 p s DGD Off set #. ~. .I§' 'WI!: -it-· - ~-.. ......... . ~Mfi" ~~ Fig. 4.9. Experimental results. BER characteristic of 10.7 and 12-Gbaud 16-QAM signals as a function of received power before and after DGD offsets; (insets) Constellation diagram of 12-Gbaud 16-QAM signals with and without DGD offset. after applying DGD offset of almost 33 ps at 12-Gbaud. Figure 4.10 shows the power penalty at 10· 3 BER for the generated 20-Gbaud 16-QAM versus DGD offset. As it can be seen, when the DGD offset is larger than 20 percent of the symbol period (e.g., 10 ps for 20-Gbaud signal), the power penalty increases exponentially. ii)15~~~~~~~~~~~ ~ 0:: w Ill ..., 10 w @ ~ ~ 5 Cl> a.. ~ 0 a.. OL__~~~~~~~~~--' ·20 ·10 0 10 20 DGD Offset (ps) Fig. 4.10. Experimental power penalty characteristic of a 20-Gbaud 16-QAM signal versus DGD offset. 44 The experimental results are a proof of concept for our polarization-based method for 16-QAM generation. In addition. it provides a simple passive scheme for emulating 16-QAM for lab-based optical signal processing applications. Another advantage of the proposed scheme is its cascadibility potential. For example. by cascading two DGD stages. we can potentially generate 256-QAM. 45 46 Chapter 5: High Speed Reconfigurable Nonlinearity Based Optical Multiplexing Technique for QAM Signal Generation In this chapter we describe a new nonlinearity-based technique for optical multiplexing of several individual amplitude modulated signals at different wavelengths into one higher-level modulated signal. We also experimentally demonstrate tunable phase- coherent optical multiplexing of four 20-Gbaud OOK signals at different wavelengths into a single 80-Gbit/s 16-QAM channel and also a single star 16-QAM channel based on coherent addition of input signal amplitudes using periodically-poled lithium niobate (PPLN) waveguides. We also demonstrate tunable phase-coherent multiplexing of two 20-Gbaud OOK signals into a single QPSK and also a single four-level amplitude/phase modulated ( 4 PAM) channel. 5.1 Introduction As described in chapter 2, higher-order modulation formats are of extreme interest to the optical communications community due to the higher spectral efficiency and higher 47 tolerance to fiber-based impairments. In particular. there is interest in 16 quadrature amplitude-modulation (QAM) generation. Typical methods for generating 16-QAM include: (a) electrical 16-QAM generation using an I-Q modulator. which can be synthesized by a pulse pattern generator (PPG) or an arbitrary wave form generator (AWG). (b) optical 16-QAM generation by combining two quadrature-phase-shift-keying (QPSK) signals with different amplitude levels in an optical interferometer. and (c) 16- QAM generation using phase-stabilized fiber interferometer [62-64]. These approaches tend not to scale easily to higher baud rates or constellation sizes. It is desirable to generate 16-QAM and higher-order modulation formats using fully optical approaches. such that several lower-level modulation formats are multiplexed into one higher-level modulation format signal. The multiplexing process is bit-rate tunable. and there is a potential for much higher baud rates and constellation sizes. In this chapter. we experimentally demonstrate tunable phase-coherent optical multiplexing of four 20- Gbaud OOK signals at different wavelengths into a single 80-Gbit/s 16-QAM channel and also a single star 16-QAM channel based on coherent addition of input signals using periodically-poled lithium niobate (PPLN) waveguides. We also demonstrate tunable phase-coherent multiplexing of two 20-Gbaud OOK signals into a single QPSK and also a single four-level amplitude/phase modulated (4-PAM) channel. respectively. 5.2 Concept Mon-off keying (OOK) input signals can be multiplexed into a 2M QAM signal based on coherent addition of initial signals. A block diagram of the proposed scheme is illustrated in Fig. 5.1. As can be seen, four on-off keying (OOK) input signals can be multiplexed 48 into a 16-QAM signal based on coherent addition of initial signals. Each point of the 16- QAM constellation in the I/Q plane can be interpreted as a coherent vector addition of one point from OOK 1, a second point from OOK 2, a third point from OOK 3, and a fourth point from OOK4. Note that the vector magnitude of OOK 1 and OOK 3 should be twice (6-dB power difference) that of the OOK 2 and OOK 4 magnitude respectively. OOKl and OOK2 are in phase, while OOK3 and OOK4 are in phase also, but with n/2 phase difference compared to OOKl and OOK2. Amplitude Optical Multiplexer ~ c 0 Amplilul,°°~~2: JI 11 : ~ 0.5 q 16-QAM Amplitude time _!!_ 1 t ~K~ c 'A- 2 W1.lW.L+ c::::::} 7r Amplitud OOK( 4 ) time ~ (/!4 =2 0.5 ..___ v time Fig. 5.1. Block diagram of optical multiplexing of 4 OOK channels into a 16-QAM signal. Figure 5.2 depicts the conceptual diagram of multiplexing of four OOK inputs at m 5 I frequencies into a 16-QAM and a star 16-QAM signal based on nonlinearity effects in PPLN waveguides. Phase-conjugated copies of incoming OOK signals are generated at 2mP - m 5 frequencies based on a cascaded second harmonic generation/ difference I 49 frequency generation ( cSHG/DFG) effect in a periodically poled lithium niobate waveguide (PPLN-1) with a continuous pump signal with frequency of cuPset near the quasi phase matching (QPM) frequency of a PPLN waveguide. The relation between generated phase-conjugated complex amplitudes and initial signal complex amplitudes is given by A; oc A/ A 5 'oc A/IAs I exp(- j(/Js ) k k k k (5-1) where A 5 is the initial signal complex amplitude. AP is the continuous pump complex ' amplitude and A;, is the phase-conjugated signal complex amplitude. and k is the number of initial OOK signals. Incoming signals are amplitude modulated signals and exp(j¢ 5 )is a random phase carried by each initial OOK signal. Note that these random ' phases must be removed in order to multiplex incoming signals into a higher-order QAM signal. After applying required phase and amplitude weights to each OOK signal using an optical waveshaper such as the Finisar 4000E. the phase-conjugated copies and the initial OOK signals are filtered and sent to a second PPLN waveguide with a similar quasi- phase matching (QPM) wavelength along with a continuous pump signal with frequency of cur. Note that all initial OOK signals become phase-coherent and are multiplexed into a unique frequency 2cuP as a result of the sum frequency generation effect (SFG) between the input OOKs and their phase conjugated copies. because this process erases the phase differences (random phases) between incoming OOK signals. The resulted multiplexed idler is out of the C band frequency domain. but is returned to the C band by difference frequency generation (DFG) nonlinear effect into a single frequency of 2cuP - cur to generate a QAM signal. The relation between the initial signal complex amplitudes and so the generated new idler complex amplitude as a result of a cascaded sum frequency generation (SFG)/difference frequency generation (DFG) is given by 4 exp(jq> 5 )exp(- jq> 5 ) oc L wk exp(jak)Ar'IAs,1 2 k=l (5-2) where wk exp(jak)) are the amplitude and phase adjustments applied to the incoming signals after cSHG/DFG process using an optical waveshaper (Finisar 4000E). Here Ar is the pump complex amplitude and A.~ictI~ is the generated idler complex amplitude as a result of cSFG/DFG process. The required amplitude and phase adjustments to generate a square 16-QAM are given by a 3 = a 4 = 90" a 1 =a 2 =0°. (5-3) The offset from generated QAM signal constellation is removed by adding a CW pump with a proper phase and amplitude. 51 Generation of Phase Conjugate Amplitude/Phase Adjustments Copies [ -T~a_p_s~1~a_,~a2..._a~}_a_,..._~~-w_._ 1 w~2 -w~ 3-w_ 4 ....._J , _ o· o· 90· oo· 1 112 1 112 , W1W2W3W4 l & a.r1 1 · QPM Fig. 5.2. Conceptual diagram of optical multiplexing of four OOK inputs into a 16-QAM. 5.3. Experimental Setup and Results Figure 5.3 shows the experimental setup. Four continuous-wave (CW) lasers are modulated to generate 20-Gbit/s OOK signals using a Mach-Zehnder modulator (MZM) Coherent Receiver Data PPLN-1 L------ Fig. 5.3. Experimental setup. CW: continuous wave; MZM: Mach-Zehnder modulator; BPF: band pass filter; PPLN: periodically poled lithium niobate; TDL: tunable delay line. 52 and are subsequently decorrelated usmg three tunable delay lines (TDL). The four independent OOK signals and a CW pump set near the QPM wavelength are then sent to the first PPLN with a QPM wavelength -1550.7 nm to generate phase-conjugated copies of OOK signals. The phase-conjugated copies of inputs and input OOK signals are then filtered and sent to the second PPLN with the same QPM wavelength after passing through a Finisar optical wave shaper for required phase and amplitude adjustments. Another CW pump is also sent to the second PPLN. The coherently multiplexed QAM signal is filtered out and detected by a coherent receiver. Back-to-back optical constellation and eye diagram of the two incoming 20 Gbit/s OOK signals are shown in Fig. 5.4 (a) and 5.4 (b). Figure 5.4 (c) shows constellation and eye diagrams of a 4-PAM O Gbitls 4-level phase/amplitude (4-PAM) 10 Gbit/s input OOKs (c) (a) (b) l Amplitude/Phase Adjustments 20 Gbitls QPSK constellation • • (d) ,... • ., ~ 0 Amplitude/Phase Adjustments 0 [ Tapsl a1 a2 W1W2 I 1 1 o· so· Fig. 5.4. (a-b) Back-to-back optical eye and constellation diagrams of two 20 Gb/s OOK inputs; (c) optical eye and constellation diagrams of a multiplexed 4-PAM signal; (d) optical constellation diagram of a multiplexed QPSK signal with EVM of 18.5. 53 (phase/amplitude modulation format) signal generated by multiplexing of two in-phase OOKs with amplitude ratio of two. Fi~e 5.4 (d) shows a constellation diagram of a QPSK signal generated by multiplexing of two OOKs with the same amplitude and n/ 2 phase difference . • ' ' , 6 • @ ' ' 0 -20 · 4 0 -60 ·80 -100 0 ·20 (a) II HS40 (b) DC emover Pump 1 e1'4 e1 1 e1!50 1ee15 Wovulcns1h (nm) DC remover Pum : e ~o .~ 60 ! -~: .ii~_._,.___.__,.__.__ ' 1540 1 545 1 550 1 555 1 560 ' wavelength (nm) ' Fig. 5.5. (a) Optical spectrum at the output of PPLNl with two incoming OOK signals and a CW pump to remove DC from the constellation; (b) Optical spectrum at the output of PPLN2. Note that the generated QPSK has an error vector magnitude (EVM) equal to 18.5. Fi~e 5.5 (a) and 5.5 (b) show the optical spectra of a 4-PAM (phase/amplitude modulation format) signal generation process at the PPLNl output and the PPLN2 output based on multiplexing of two in-phase OOKs with an amplitude ratio of two. A pump with proper phase and amplitude is also sent to PPLNs along with initial OOKs to remove the offset DC value from the 4-PAM constellation. Fi~res 5.6 (a) and 5.6 (b) 54 show the optical spectra of a 16-QAM signal generation process at the PPLNl output and the PPLN2 output by multiplexing of four OOKs with the required phase and amplitude as described in section 5.2. Figure 5.6 (a) shows the generated idlers after cSHG/DFG in PPLN 1 while Fig. 5 .6 (b) shows the generated idler result ( 16-QAM multiplexed signal ) after cSFG/DFG in PPLN2. (a) -20 E -40 = -.::::> -60 -80 (b) -20 -40 E (() -0 -60 -80 -100 4-00Ks 4- converted OOKs 1540 1545 1550 1555 1560 \/\/avelen!'.]th (nm ) Optical spectrum al the output of PPLN 2 (no DC offset remover) Off set 1 6-QAM 1540 1545 1550 1555 1560 Wavelength (nm) Fig. 5.6. (a) Optical spectrum at the output of PPLNl with four incoming OOK signals; (b) Optical spectrum at the output of PPLN2. Note the cSHG/DFG effect in PPLNl is a process to generate phase-conjugated copies of initial OOKs. Cascaded SFG/DFG nonlinear effect in PPLN2 can remove random phases between the incoming signals by adding the original OOKs and their phase-conjugated copies to make them phase coherent and return all OOKs to a single wavelength to generaLe a single wavelength higher-order multiplexed modulation format signal. 55 Optical spectrum at the output of PPLN 1 (with a DC remover pump) (a) 0 -20 E -40 .g -60 -80 -100 (b) 0 -20 Optical spectrum at the output of PPLN 2 (with a DC r E -40 ig -60 -80 -100 -. 5~5 -i.550 1 555 W.:c.......,..E!tloet'..,,:Q'""t::h· Cr.m') Fig. 5.7. (a) Optical spectrum at the output of PPLNl with four incoming OOK signals and a continuous pump as a DC offset remover; (b) Optical spectrum at the output of PPLN2. Figures 5.7 (a) and 5.7 (b) show the optical spectra of a 16-QAM signal generation process at the PPLNl output and the PPLN2 output by multiplexing of four OOKs with the required phase and amplitude as described in section 5.2. A pump with proper phase and amplitude is also sent to PPLNs along with initial OOKs to remove the offset DC 40 Gbit/s multiplexed16- QAM signal .. ~ t: ....... . . ,... .... . ........ ... . •" Jt 40 Gbit/s multiplexed Star-16QAM signal Fig. 5.8. Optical constellation diagram of 40 Gbit/s square and star 16-QAM signals generated by multiplexing of four 10 Gbit/s OOKs. 56 80 Gbit/s multiplexed 16-QAM signal 80 Gbit/s multiplexed Star-16QAM signal Fig. 5.9. Optical ronstellation diagram of 80 Gbitfs square and star 16-QAM signals generated by multi}iexing of four 20 Gbitfs OOKs. value from 16-QAM constellation Shown in Fig. 5.8 arrl Fig. 5.9 are the optical constellation diagrams of 40 Gbitis and 80Gbitis square and star 16-QAM generated by multi}ixing of four 10 Gbitis arrl 20Gbitfs OOK signals. Shown in Fig. 5.10 are the optical constellation diagrams of 30 Gbitis rectangular and diagonal 8-PAM signals generated by multi}iexing of three OOK signals. 30 Gbit!s multiplexed rectangular and diagonal 8-PAM signals 0 0 ...... • · •= ~- 0 - - • 0 • .. 0 - . ,: . •: 0 . 0 ... 0 ~- -- 0 C#Jt- 0 0 • Fig. 5.10. Optical constellation diagram of 30 Gbitis rectangular and diagonal 8-PAM signals generated by multiplexing of three 10 Gbitis OOKs. 57 5.4. Further Discussion: Optical Multiplexing of a 16-QAM and a QPSK into a 64-QAM Signal Multiplexing of the lower-level modulation formats (e.g. on off keying (OOK) signals) into a single wavelength higher-level modulation format needs a technique to eliminate the random phase differences between incoming OOK signals. As can be seen in section 5.2, the scheme that we proposed resulted in generation of an offset 16-QAM signal. That is a very useful result as it makes the path of multiplexing the offset 16-QAM signal and a QPSK signal with four phase states {o "' "' - "'} easy. Multiplexing of an offset '2' ' 2 16-QAM signal and a QPSK into a 64-QAM signal can be achieved based on cascaded sum frequency generation/difference frequency generation ( cSFG/DFG) nonlinear effect in a PPLN waveguide. The relation is given by (5-4) where ~PsK is the initial QPSK complex amplitude, ~AM is the initial offset 16-QAM complex amplitude, AP is the continuous pump complex amplitude and AM is the generated 64-QAM complex amplitude. Therefore each point of a 64-QAM constellation can be interpreted as a rotated version of one point from the offset 16-QAM constellation, and the amount of this rotation is defined by the phase of the centered QPSK constellation points. Note that the centered QPSK has four different phase states 0- ff-. { 7r - "'} '2' ' 2 58 The centered QPSK signal will map the offset 16-QAM constellation to the four different quadratures, thereby obtaining a complete 64-QAM constellation, The same concept can be applied to multiplex an offset M-QAM signal and a centered QPSK signal into a (4*M)-QAM signaL 59 60 Chapter 6: High Speed Reconfigurable Nonlinearity Based Optical Demultiplexing Techniques for QAM Signal Processing In this chapter we describe a new technique for optical demultiplexing of a single wavelength higher-level modulation format into individual lower-level modulation formats based on a method that we call nonlinearity-based phase squeezing. We demonstrate simulation results for optical demultiplexing of a QPSK signal into two BPSK sub-channels. Both sub-channels of a QPSK signal are extracted simultaneously and independently. This method potentially may be applied to higher-order QAM signals as well. We also propose a new fully-optical technique to obtain different logic and arithmetic relations between sub-channels carrying a 16-QAM signal without the need for a coherent receiver. We also demonstrate experimental results for logic and arithmetic relations between sub-channels carrying by 10 Gbaud 16-QAM signal. 61 6.1. Introduction There has been dramatic growth in capacity demand in networks, necessitating a simultaneous dramatic increase in the data speeds of transmitters and receivers. State-of- the-art transceivers are capable of providing 100-Gbit/s Ethernet data rates per channel, and these channels employ the use of spectrally efficient higher-level modulation formats as well as polarization multiplexing and coherent technologies. Bina I Binary CH3 Bina CH4 Op deal Network •••• •••• •••• •••• Coherent Receiver Binary CH 1 - BinaryCH2 - Binary CH 3 - Binary CH 4 - ., CH 1(0R)CH2 - Logic& CH 1 (NANO) CH2 Arlthmedc - Gate CH 1(NOR) CH2 - CH 1(+) CH2 , - Fig. 6.1. Principle of the conventional method to extract sub-channels from a QAM signal using an electrical coherent receiver. Although ultra-high-performance transceivers provide high capacity data pipes, there is a large discrepancy between high-rate and low-rate data channels, so that the large capacity of a single data channel from the transceiver is often not required and is under-utilized. A laudable goal is to create transceivers whose large capacity can be tailored and shared among many different channels by modifying the bit rate and modulation formats as the traffic demands vary in a dynamic network. The optical demultiplexing of higher-level modulation format signals into lower level modulation format sub-channels using fully optical modules has a potential impact on reconfigurability of future optical networks. 62 Typical methods for demultiplexing of higher-level modulation format signals include: (a) optical DQPSK format conversion [69-71], and (b) optical sub-DQPSK channel information extraction from a D8PSK signal [72]. A laudable goal is a scalable optical method to demultiplex quadrature/ amplitude modulation (QAM) signals into lower-level modulation format sub-channels without missing any sub-channels. In this chapter. we describe a scalable optical technique to demultiplex QPSK data channel into two BPSK sub-channels using a nonlinearity-based scheme. This method potentially may also be scalable to demultiplex higher-order QAM signals as well. Figure 6.1 shows the principle of a conventional method to extract sub-channels from a QAM signal using an electrical coherent receiver. As can be seen. even to obtain the information related to different logic and arithmetic functions between sub-channels carrying by a 16-QAM signal. the higher level QAM signal is first demultiplexed completely by an electrical coherent receiver. Our goal is to perform this post-processing function using a fully-optical module without requiring an electrical coherent receiver. Fully-optical signal processing modules can avoid optical-electronic conversion inefficiencies and take advantage of large bandwidth and wide range of tunability in the optical domain. In this chapter. we describe a scalable optical technique to obtain different logic and arithmetic functions between sub-channels carrying by a 16-QAM signal without the need to fully demultiplex the original QAM signal. 6.2. Phase Sensitive Amplification Concept The principle of phase sensitive amplification is shown in Fig. 6.2. In a phase insensitive amplifier such as an erbium-doped fiber amplifier (EDFA) as shown in Fig. 6.2 (a), the 63 in-phase and quadrature electric field components are amplified equally, and consequently the electric field amplitude is amplified but the phase does not change. However, as shown in Fig. 6.2 (b) in phase sensitive amplifiers (PSAs), the in-phase and quadrature components are amplified differently. For example, in a phase sensitive amplifier based on degenerate four-wave mixing (FWM), the in-phase component of the electric field is amplified by a gain factor of y ea;n, while the quadrature component of the electric field is attenuated by a factor of aA 11 [73-74]. As a result, the output phase of the amplified electric field is closely aligned with the in-phase axis of the amplifier. This process is called 'phase-squeezing'. Phase insensitive amplification Phase sensitive amplification (a) (b) Im ( quadr ab.Jre) Fig. 6.2. Principle of (a) phase insensitive and (b) phase sensitive amplifiers. 6.3. QPSK Signal Demultiplexing Concept Figure 6.3 shows a conceptual diagram of a QPSK signal demultiplexing based on phase sensitive amplification. In phase sensitive amplification (PSA) process, the signal is added coherently to its phase-conjugated copy. The first BPSK sub-channel is obtained if 64 the signal phase is squeezed along the imaginary axis while the second BPSK sub- channel is obtained if the signal phase is squeezed along the real axis. Note that a PSA configuration along the imaginary axis is obtained by applying a 'lf /2 phase shift on the initial signal. As shown in Fig. 6.3 (a), phase squeezing along the imaginary axis results in mapping the constellation points carrying the same first bit to a single point, corresponding to recovering BPSK 1 sub-channel constellation. Also as shown in Fig. 6.3 (b ), phase squeezing along the real axis results in mapping the constellation points carrying the same second bit to a single point, corresponding to recovering BPSK 2 sub-channel constellations. The phase squeezing technique is implemented through coherent nonlinear wave mixing in a HNLF or a PPLN waveguide [73-7 4]. Initial QPSK 10 00 • 0 0 0 11 01 + -------, I Phase conjugated l copy of QPSK data I ------· Phase Sensitive Amplifier( along x axis) 10 • 11 r -->Thesame first bit • (a) BPSK1 ~---·1 ---40.:::>+ BPSK2 ~ :t Phase Sensitive Amplifier( along y axis) 0 ~ J§--+ The same second bit __ _ .. I' (b) Fig. 6.3. Principle for demultiplexing of a QPSK into BPSK sub-channels based on phase sensitive amplification. The two BPSK sub-channels at the output of the optical module may be sent to two different nodes in a network without any optical-electronic-optical conversion. The 65 relation between input signal electric field and output signal electric field in a phase sensitive amplifier module under certain conditions can be given by [73-74] Re(E 0 u,) = Re(E;n)·Yaa;n (6-1) 6.4. Results and Discussions Figures 6.4 (a-d) show demultiplexed BPSK sub-channels constellation diagrams simulated in Matlab for different gain factors. The phase squeezing effect requires a high ratio between the gain factor along the real axis and the attenuation factor along the imaginary axis. Figure 6.4 (a) depicts the demultiplexed BPSK sub-channels constellation diagrams with a gain factor of 3 .5 dB. As can be seen the phase-squeezing effect is not strong enough and constellation points are still separated. Figure 6.4 (b) shows the demultiplexed BPSK sub-channel constellation diagrams for a module with a gain factor of 6 dB. With 6 dB gain, the phase squeezing effect is yet not strong enough and the BPSK sub-channels are not extracted without error. Shown in Fig. 6.4 (c) and Fig. 6.4 (d) are the demultiplexed BPSK sub-channel constellation diagrams for a module with a gain factor of 9.5 and 12 dB. As can be seen from the constellation diagrams the BPSK sub-channels are extracted and the BER is less than 10- 4 • 66 (a) BPSK sub-channels Gain factor= 3.5dB (b) BPSK sub-channels Gain factor= 6dB ------ ------ ------- ------ ' \ I \ I \ I \ I • • I I · ·· ·· I I · · I I · : · · : ·· I . c · z ii •~ ·• ! ii T: + :: ~r~ : ' .l. - ~ - :_ ~ ' - - - - - " (c) BPSK sub-channels Gain factor= 9.5dB ------llllllt ~------ ( ; .. ; \ ( \ • .. ···• I ...... . •...... , . , . .. I I . • . '.· . : I f , • · : . . .• • I 1··· · ······ ·· I ' ~--~···~--~ ... ~ ,; ' -----,; , _____ ; '-----"' (d) BPSK sub-channels Gain factor= 12dB ------ ------ ' \ I \ I • 11 I I 11 I • I I I I .. . : . . I I jj JI I ' -. ..... ... _ .... ; ' ----- ; Fig. 6.4. (a-d) Constellation diagrams of demultiplexed BPSK sub-channels for different gain factors. Figure 6.5 shows the constellation diagram and BER vs. optical signal-to-noise ratio (OSNR) for the original QPSK signal as well as BER vs. OSNR for the two demultiplexed sub-channels with a gain factor of 9.5 dB. As can be seen for BER of 10- 4 , the OSNR value is almost lOdB. Note that at BER of 10- 4 the OSNR difference between the demultiplexed sub-channels and initial QPSK channel is less than 1 dB. The phase squeezing technique may also be applied to higher-level modulation formats (e.g., 16-QAM). However, achieving the optimum gain factor, as well as the technique to generate phase locked pumps becomes more challenging. 67 10.J ~ fol ~ 10 4 ';" 1: 1 channel ,_ 10-5 7 8 9 10 11 12 13 14 OSNR( dB) Fig. 6.5. BER vs. OSNR of the original QPSK signal and the two demultiplexed channels with gain factor of 9.5 dB; Inset: Constellation diagram of the original QPSK signal. 6.5. All-Optical Logic/Arithmetic Gate Implementation for QAM Sub Channels In this section, we describe an optical module that produces various forms of logic and arithmetic relations between sub-channels carried by a 16-QAM signal based on the four wave mixing effect in highly nonlinear fibers (HNLF) without the need to fully demultiplex a 16-QAM signal. Other potential applications of this fully-optical module might be in the area of optical switching. 6.5.1. Concept Figure 6.6(a) depicts a conceptual diagram of 16-QAM generation. Each point of a 16- QAM constellation in the I/Q plane can be interpreted as a rotated version of one point 68 from the offset 4-QAM (QPSK) constellation, and the amount of this rotation is defined by the phase of the centered QPSK constellation points. As is shown by different color codes, the centered QPSK signal will map the offset QPSK constellation points to four (a) Offset QPSK 1 symbols (AB) Square 16 QAM ooeo Optical Phase ErasureModule Square 16 QAM OOeo eooo 000• oeoo 00 Threshold 1 A(OR)B, A(NOR)B (b) • O O O ....... Signal + 0 0 0 • ...,,.. Phase Conjugated Cop oeoo Threshold 2 A(AND)B, A(NAND)B, (A*B) ( <p) + (-<p) = 0 QPSK1 symbol information Fig. 6.6. (a) conceptual diagram of 16-QAM generation; (b) Principle of all optical logic/arithmetic gates implementation for 16-QAM sub-channels. different quadratures, thereby obtaining complete 16-QAM constellations. Shown in Fig. 6.6 (b) is a simple optical approach that performs as a phase erasure for 16-QAM signal based on the four wave mixing (FWM) effect using two cascaded HNLF stages. The relation between the phase-conjugated complex amplitude and initial signal complex amplitude is given by A' OC A 2 Pump 1 As* OC A 2 Pump 1 IAslexp(-)</Js) (6-2) where A 5 is the initial signal complex amplitude, A' is the phase-conjugated signal complex amplitude, and APu11p 1 is the CW pump complex amplitude that is sent to HNLFl along with the original 16-QAM signal. The optical phase erasure effect occurs in the second HNLF and the relation between the initial signal complex amplitude and 69 the generated new idler complex amplitude as a result of the four wave mixing (FWM) effect is given by A" oc A* Pump,AsA' oc A* Pump, IAsl 2 exp(-l'Ps )exp(jcps) oc A* Pump, IAsl 2 (6-3) where A" defines the new idler complex amplitude as a result of the FWM effect in the second HMLF, and AP,,,,,, is the CW pump complex amplitude that is sent to HNLF2 along with the original 16-QAM signal and the phase-conjugated copy. As can be seen, all the phase states in the 16-QAM signal are erased and the resulting output is a 3-level amplitude modulated signal corresponding to the three amplitude rings of the initial 16- QAM signal. In the other word at the output of this optical module, symbol information of the offset QPSKl is extracted in the form of a 3-level amplitude modulated signal in which the lower level indicates (11) symbol information, the middle level indicates (01, 10) and the upper level of the 3-level amplitude modulated signal carries (00) symbol information of QPSKl. OR and NOR logic functions between QPSKl symbols are obtained using threshold level 1, while AND, NAND logic functions between QPSKl symbols are achieved using threshold level 2. Note that the multiplication arithmetic operation between QPSKl symbols can be obtained by picking threshold level 2, while the addition arithmetic operation can be achieved without any threshold level. 70 6.5.2. Experimental Setup The experimental setup is shown in Fig 6.7. A continuous-wave (CW) laser with 100- kHz linewidth at -1557 nm is modulated by a dual-parallel Mach-Zehnder modulator (MZM) which is driven by two 2 11 -1 pseudo-random bit sequence (PRBS) to generate a QAM Generator Data I EDFA TDL l/Q Mod : --~-,1 .......... J it-- DataQ HNLF PC Receiver HNLF PC Fig. 6.7. Experimental setup. CW: continuous wave laser: PC: polarization controller: BPF: band pass filter: TDL: tunable delay line: HNLF: highly nonlinear fiber. 10-Gbaud QPSK signal. The generated QPSK signal is then launched into an integrated 16-QAM emulator consisting of an optical splitter, a two bit delay and 6-dB attenuation in one arm and a combiner at the output. This setup converts the QPSK signal into a 16-QAM signal which is sent to the first HNLF along with a CW pump at 1555.6 nm. The first stage of nonlinearity is a 460-m HNLF with a zero-dispersion wavelength (ZDW) of ~1556 nm and a nonlinear coefficient (y) of 20 w- 1 km- 1 • As a result of the FWM nonlinear effect, the phase-conjugated copy of original signal is generated at 1554.2nm, which is further selected by a band pass filter and is launched into the second 71 HNLF along with the original signal and a CW pump at 1553.2 run. The second nonlinear stage is a 520-m HNLF with a zero-dispersion wavelength (ZDW) of -1555 nm and a nonlinear coefficient (y) of 20 w- 1 km- 1 • As a result of non-degenerate FWM, an idler is generated at 1558 run which is filtered out by a band pass filter and sent to the receiver. 6.5.3. Experimental Results and Discussions Figure 6.8 (a) shows the eye diagram and constellation diagram of the back-to-back 16-QAM signal with EVM of approximately 8.6, while Fig. 6.8(b) shows the 16-QAM constellation diagram at the output of HNLFl with EVM of approximately 9.7 and the constellation diagram of the phase-conjugated copy at the output of HNLFl with EVM of (a) B-to-B constellation d iagram EVM - 86 .- ... ... ...... -.· .. - .-. ,,.,_ .. 31': .. Initial signal constellation at HNLF1 output (b) EVM= 9 .7 ifli" · .,,,,. . .. .. "'.: ·~ -~ ·-•· .. .... ...-. - -·· ... B -to-B eye diagram EVM- <3.6 Phase conjugated copy constellation at HNLF1 output EVM= 9 . 9 .... ,.,.. ......... ~-~-~ . • 41l· .. - ·· - ·~:, ... ~ . Fig. 6.8. (a) Optical eye and constellation diagrams of 10-Gbaud back-to-back 16-QAM. (b) constellation diagram of 10-Gbaud 16-QAM at HNLFl output, constellation diagram of the phase-conjugated copy of 10-Gbaud 16-QAM at the HNLFl output. 72 approximately 9.9. The optical spectrum at the output of HNI.Fl is shown in Fig. 6.9 (a). Here, an idler with a conversion efficiency of -16 dB is generated, which is the phase-conjugated copy of the initial 16-QAM signal. The generated idler is then selected by a band-pass filter and is sent to a second HNI.F along with the original signal and a CW pump. The optical spectrum at the output of HNI.F2 is shown in Fig. 6.9 (b). Here, an idler with the conversion efficiency of around -15 dB is obtained which is the 3- level amplitude modulated signal. Optical Spectrum After the First Noriinear Stage (a) .... ! ····+ i ·+ ptical Spectrum After the See-0nd Nonlinear stage (b) 4 i 3-amplitu:le level -. --+i. >--•-w.-t-;-H---*- signal --· .;+>--.. --... - · ! j Fig. 6.9. Optical spectrum. (a) at the output of HNLFl; (b) at the output of HNLF2. Figure 6.10 shows the BER vs. optical received power for back-to-back (B2B) 16-QAM signal, the 16-QAM signal at the output of HNLFl and the phase-conjugated copy of initial 16-QAM signal at the output of HNI.Fl. Figure 6.11 shows the EVM vs. optical received power for back-to-back (B2B) 16-QAM signal, the 16-QAM signal at the output of HNLFl and the phase-conjugated copy of initial 16-QAM signal at the output of HNI.Fl. As can be seen for BER of 10- 3 , the power penalty for the original signal at the 73 15 -r--------------1- 10 Gbaud phase 2 2 5 cc 3 w Cll -;:; 3 5 0 ...J 4 4 5 Conjugated copy of 16 -QAM signal - 10 Gbaud B2B 16-QAM signal - 10 Gbaud 16-QAM signal at the output of HNLF1 - 35 - 30 -2 5 recevied power(dbm) Fig. 6.10. BER performance versus optical received power. output of HNLFl is approximately 1 dB and for the phase-conjugated copy 1s approximately 1.4 dB in comparison to the back-to-back original signal. 14 12 '# 11 ~ 10 > w -40 -35 -30 --10 Gbaud phase Conjugated copy of 16QAM signal --10 Gbaud B2B 16 QAM signal - 10 Gbaud 16 QAM signal at the output of HNLF1 -25 recevied power( dbm) Fig. 6.11. Error vector magnitude (EVM) versus optical received power. 74 6.5.4. Experimental Results of Logic/Arithmetic Gate Implementation Between Sub-Channels Logic and arithmetic operations implemented in the optical domain can potentially enable signal processing functions at the high-speed optical network, There have been several demonstrations of logic functions based on various nonlinear processes in optical fiber, PPLN and semiconductor devices, However, as the goal of this chapter is to Eye diagram of 10 Gbaud 3-amplitude level signal 10,01 Fig, 6,12, Eye diagram of the 10-Gbaud 3-level amplitude modulated signaL address optical demultiplexing (sub-channel information recovery) of higher-order modulation formats, the proposed method to recover 16-QAM signal sub-channels is by obtaining logic/arithmetic relation between symbols, Figure 6, 12 shows the eye diagram of the 3-level amplitude modulated signal at the output of HNLF2, The summary of all-optical logic functions and arithmetic operations between QPSKl symbols is shown in table,6, L 75 Note that in many applications, logic and arithmetic functions between two signals are needed rather than the actual value of each individual signals. These may be achieved based on the proposed optical phase erasure technique for QAM signals. A B A(AND)B A(NAND)B A(OR)B A(NOR)B A+B A*B Threshold2 Threshold2 Threshold! Threshold 1 No Threshold Threshold2 0 0 0 1 0 1 0 0 0 1 0 1 1 0 1 0 1 0 0 1 1 0 1 0 1 1 1 0 1 0 2 1 Table. 6.1. All optical logic gate and arithmetic operation implementations between symbols carrying by a 16-QAM signal. 76 Chapter 7: All-Optical Phase-Preserving Multilevel Amplitude Regeneration Using Coherent Polarization Mixing In this chapter we propose a new technique for all-optical amplitude regeneration of higher-order modulation formats based on a method that we call coherent polarization mixing. We describe two different schemes for the proposed polarization-based multilevel amplitude regeneration technique. The schemes provide a wide range of tunability and scalability, and have a simpler configuration compared to previous methods. We specifically propose and demonstrate all-optical tunable phase-preserving amplitude regeneration schemes for optical star 8-quadrature-amplitude modulation (star-8QAM) signals and star-16QAM. We show that amplitude noise can be efficiently suppressed on both amplitude levels. The system robustness against nonlinear phase noise originating from the Gordon-Mollenauer effect in a 150 km transmission line is investigated using the proposed amplitude regenerator. 77 7.1 Introduction There has been dramatic growth in capacity demand in networks, necessitating a simultaneous dramatic increase in the data speeds of terminal transmitters and receivers [75-76]. As can be seen in Fig. 7.1 the main sources of amplitude noise are amplified spontaneous emission (ASE) added to the optical pulses at each in-line EDF A and nonlinear intra- and inter- channel effects [77-79]. Moreover, for phase modulated formats, amplitude noise can be converted into nonlinear phase noise by nonlinear effects such as the Gordon-Mollenauer effect in optical fiber transmission lines [80-81]. All- optical regenerators are expected to extend the transmission maximum reach of high- (a) , •• ••• 1 .. I 1{ -.~ Optlcal .. Regenerator Orlglnal Pulse Fiber Degraded Pulse (b) +i 0 riginal signal constellation diagram + Degraded signal constellation diagram Amplltled Pulse q Amplifi d signal constellation diagram Regenerated Pulse +i Regener a too signal constellation diagram Fig. 7.1. (a) An optical pulse degraded by amplitude noise in an optical fiber transmission line. (b) Constellation diagrams of an optical signal at different points in an optical transmission line. speed transmission systems by eliminating accumulated signal impairments in transmission systems without the need of optical/electronic/optical (O/E/O) conversion. Because amplitude fluctuations can also be converted into phase noise due to fiber 78 nonlinearities, phase-preserving amplitude regeneration is a desirable function for phase- modulated signals [82]. Typical methods for amplitude regeneration of phase modulated signals include: (a) using optical limiters, which can be achieved by optical parametric amplification (OPA) based on four wave mixing [83]; (b) using amplifying loop mirror [84 ]; and ( c) using cross phase modulation effect and offset filtering [85]. Fiber parametric amplifiers (FPAs) have been studied previously for amplitude regeneration of OOK signals [86-87] and wavelength conversion of DPSK signals [88]. Theoretical analysis shows their potential to regenerate the amplitude of DPSK signals without significantly degrading phase information [89]. Recent experiments have confirmed both limiting amplification of pulses with small phase noise introduction [90], pump FSDL ~ DD DPSKDM signal PC monitor Fig. 7.2. Experimental setup. FSDL: free space delay line; PC: polarization controller; DD: direct detection; DPSK DM: demodulation using delayed interferometer; OSNR: optical SNR adjustment by adding ASE [83]. and the reduction of accumulated nonlinear phase noise achieved using a regenerator before a fiber transmission span [91]. Figure 7.2 demonstrates the experimental setup for FPA-based amplitude regenerator for DPSK signals. Light from independent tunable laser sources ( A.s = 1549.5 nm, A.P = 1552 nm) is modulated with data streams of PRBS 2 15 -1 at 10 Gbit/s to generate a RZ-DPSK 79 signal. One branch of the generated signal is decorrelated using a free space delay line and is used to suppress stimulated Brillouin scattering (SBS) in the HNLF. The pump is amplified using the EDFA and filtered to suppress amplified spontaneous emission (ASE). The pump is sent into the HNLF through a 3 dB coupler at up to 21.5 dBm peak power. Note that the HNLF length, nonlinear coefficient, loss and zero-dispersion wavelength are 6 km, 11 W/km, 0.8 dB/km and 1552 nm respectively. An EDFA is used to boost the signal launch power to 14.2 dBm peak power. The signal output is filtered and divided for direct detection and DPSK demodulation. Figure 7.3 demonstrates the power transfer function of a FPA-based regenerator at a pumping power of 21.5 dBm. As can be seen a plateau region occurs for signal powers greater than 8 dBm. Note that the signal power can be maximized in order to reduce the required pump power for gain saturation. 16 12 E 8 co "'O a..- 4 0:: 0 -4 -4 0 4 8 12 LP, dBm Fig. 7.3. Power transfer function of FPA-based amplitude regenerator (LP: signal launched average power; RP: received average power) [83]. Figure 7.4 (a) shows a noisy signal intensity eye diagram. The noisy signal was launched at maximum power co-polarized with the pump. Figure 7.4 ( c) shows the regenerated 80 I I ' I • ' ' ' I I --t----... ---- : l • I . --4--'--4---- -~--r--'"1--r-- • • I • --'---.J--"-- -._,,-.-~--r---t- 1 I I I __ .__4 __ ..__4_ • I • I I ' t I • I I I I I I • I _,.._., ___ _ I -4-- . J_ __ - a c I ' I . ____ .J_ ... _.J __ I I I I I l ~ I I I t . ~.._tll";.#' --r----,--- -1-·r-,-- _l ___ _l_ ___ L~-~--L-~- b d Fig. 7.4. (a-c) Intensity eye diagrams before and after amplitude regeneration; (b-d) demodulated eye diagrams before and after regeneration [83]. signal intensity eye diagram. The output normalized intensity noise was reduced by adjusting the pump power such that average signal peak power was sufficient to saturate gain. The corresponding demodulated eye diagram before and after amplitude regeneration is shown in Fig. 7.4 (b-d). ON-OFF keying amplitude regeneration has been achieved using interference in a nonlinear optical loop mirror (NOLM) configuration [92-94]. The NOLM-based regenerator uses a simple setup, consisting of an amplifier, a coupler, and a piece of fiber. To adjust the NOLM to regenerate optical phase encoded signals, the phase information must be taken into account. The key element is a directional attenuator/amplifier which is inserted into the conventional NOLM setup. Signal regeneration by a NOLM is based on constructive and destructive interference [92-94]. 81 1 a) e) 0 i.-:::....-- 0 0.5 1 0 0.5 1 0 0.5 I Time [l/40GHz] Time [I /40GHz] Time [1 /40GHz] Fig. 7.5. Optical eye diagrams: (a) input signal with 7% amplitude fluctuations; (b) regenerated after a conventional NOLM; (c) demodulated RZ-DPSK signal after a conventional NOLM; (d) Regenerated by the modified NOLM (e) demodulated RZ DPSK signal regenerated by the modified NOLM [84]. Figure 7.5 (a) shows the eye diagram of a noisy DPSK signal with 7% amplitude fluctuations. As shown in Fig. 7.5 (b), the amplitude jitter of the optical DPSK is reduced after conventional NOLM-based amplitude regeneration. However from Fig. 7.5 (c), after demodulation, ghost pulses and amplitude jitter appears because either the phase difference of the pulses is no longer exactly zero, or due to the induced phase fluctuations [96]. As shown in Fig. 7.5 (d), the amplitude jitter of the optical DPSK is reduced after modified NOLM-based amplitude regeneration. The optical demodulated RZ-DPSK signal in Fig. 7.5 (e) shows no degradation, demonstrating the phase-preserving characteristic of modified NOLM-based amplitude regeneration technique. 82 ~ ~ 2.5 ...=.. ~ 2.0 ~ 0 1.5 0.. ~ 1.0 ~ 0.5 .... -.... .. .. .. . . -.. -. 0.05 0.0~ ¢::< -0.05 :a en 0 -- .. ~ -0.10 ~ - ...... - - • • i::Loe . -0.15 0 0"""""=----'---~--'-~----'-~~L._~-'---~--'-~----'-~~L.._~-'---~-' 0 100 200 300 400 500 600 700 800 900 1 000 NOLM input power [mW] Fig. 7.6. Power transfer function of NOLM-based amplitude regenerator; Output power and phase shift versus NOLM input power [84]. Figure 7.6 demonstrates the power transfer function of a modified NOLM-based regenerator. Here, a plateau region exists for signal powers greater than 450 mW and less than 800 mW. At about 650-mW peak pulse power, amplitude jitter suppression is maximized, which is the center of the plateau region. The characteristic curve of the phase shows a plateau region at about 650 mW as well. Therefore, amplitude jitter causes only low excess phase shift. Note that the existing phase noise at the NOLM input is not reduced by this setup. Note that NOLM-based amplitude regenerator can be modified to reduce amplitude noise of multilevel modulation format signals [95). Among the phase-preserving amplitude regeneration techniques, those based on semiconductor optical amplifiers (SOAs) have advantages of compactness and low operating power levels. The cross-phase modulation (XPM) effect in an SOA followed by offset optical filtering has been shown for multi-channel regeneration of RZ-DPSK data [97]. All-optical phase-preserving amplitude-regeneration of NRZ-DPSK signals has been demonstrated in a saturated SOA [98-99). It is based on the amplitude-limiting property of a saturated SOA and small amplitude-to-phase noise transfer [100-101]. All- 83 optical phase-preserving amplitude-regeneration of 40 Gb/s DPSK has been demonstrated in a single SOA [138]. Figure 7.7 shows the experimental setup for all-optical phase-preserving amplitude regeneration of 40 Gb/s NRZ-DPSK signals. PC: \ "OA LD ~ ;.(W EDF'A n '1u11 Snmplutfl Scope BER te:ite1 Fig. 7.7. Experimental setup for all-optical regeneration and regenerative wavelength conversion of 40 Gb/s DPSK data (LD: laser diode; MZM: Mach-Zehnder Modulator; BPG: bit pattern generator; PC: polarization controller; VOA: variable optical attenuator; EDFA: erbium-doped fiber amplifier; OF: optical filter; SOA: semiconductor optical amplifier; DI: delay interferometer; BER: bit-error rate) [138]. Two laser diodes, LD 1 and LD2, provide the signal at 1549 nm and CW pump at 1547 nm, respectively. The output of LDl is modulated with a 40 Gb/s 2 31 -1 PRBS by a Mach-Zehnder modulator (MZM) biased at a null point to generate 40 Gb/s NRZ-DPSK data. The modulator output is coupled with the ASE generated by the EDFA and filtered by a bandpass optical filter (OF) with 1 nm bandwidth centered at 1549 nm. A variable optical attenuator (VOA) is used to control the ASE light level to adjust the OSNR of data signal. The noisy signal and the CW pump are then coupled and sent into a multi- quantum-well SOA with nearly 27 dB small-signal gain, 13 dBm output saturation power, and <1 dB polarization-dependent gain [138]. The SOA driving current is 320 mA. 84 CW 40 Gb!s DPSKData ~4 U46 1 l48 lSJO UJ! \\'a•·elenp tb (nm) Fig. 7.8. Typical spectrum at SOA output. FWM conversion efficiency is about -13 dB [138]. The typical SOA output spectrum is shown in Fig. 7.8. As can be seen, a FWM component is generated at 1545 run. Either the generated new component or the original signal can be selected at the amplifier's output by means of a tunable optical filter with 1 nm bandwidth, and is delivered to a 40 Gb/s DPSK receiver including a 25-ps delay- interlerometer (DI) demodulator. Figure 7.9 shows the input/output demodulated eye diagrams. The corresponding demodulated output eye diagrams show an improved Q-factor (with respect to input signal) of 7.7 and 9.1 at the pass-through (PT) and FWM wavelengths, respectively. 5 ps dh -. 5 ps dn-. 5 ps dJr. Fig. 7.9. Input and output eye diagrams for four-wave-mixing (FWM) and pass-through (PT) signals when input Q-factor is set to 5 [138]. 85 The BER versus receiver threshold margin is shown in Fig. 7.10. Here, the floor for the input signal at a BER value of~ 10- 10 is removed for both the PT and the FWM data. LE-3 lE-4 o IN A FWJ\l IE-5 o PT 1£-6 c:: t.;.J IE-7 a:i lE-S 1£-9 IE-10 IE-II -~O -30 -20 -10 0 10 20 30 40 Nom1alized Threshold (mV) Fig. 7.10. BER versus threshold margin for four-wave-mixing (FWM) and pass-through (PT) signals when input Q-factor is set to 5 [138]. The amount of ASE light applied on the input data has then been increased, and the input Q-factor reduced to 4. Figure 7.11 shows the input/output demodulated eye diagrams. The corresponding demodulated output eye diagrams show an improved Q-factor (with respect to input signal) of 6.3 and 7.6 at the pass-thorough (PT) and FWM wavelengths, respective! y. 5 ps d11: 5 /JS dh: 5 ps dn: Fig. 7.11. Input and output eye diagrams for four-wave-mixing (FWM) and pass-through (PT) signals when input Q-factor is set to 4 [138]. 86 The BER versus receiver threshold margin for the input Q-factor of 4 is shown in Fig. 7.12. In this case the floor in the BER could not be removed but the output signals at both the original and FWM wavelengths are improved compared to the input signal. Note that the FWM component shows a larger threshold margin improvement in the measured BER compared to the PT signal [138]. lE-3 IE-4 1£-5 ~ w co IE-6 IE-1 a IN Ii FWM lE-8 o PT -30 -20 -10 0 lO 20 30 Nonnalized Threshold (mV) Fig. 7.12. BER versus threshold margin for four-wave-mixing (FWM) and pass-through (PT) signals when input Q-factor is set to 4 [138]. In general, these amplitude regeneration techniques tend to be limited to a single amplitude level. While the NOLM-based method has been used for multilevel amplitude level signal regeneration [136-139], NOLM-based techniques suffer from disadvantages such as instability and the need for special fibers, amplifier and attenuators for high performance. A desirable goal is to implement amplitude regeneration of higher-order modulation format signals in a tunable fashion such that variable levels of amplitude noise can be accommodated using simpler schemes. In this chapter, we demonstrate 10 Gbaud optical amplitude regeneration of star QAM signals based on coherent polarization mixing using a polarizer. Two different schemes are described, and these 87 tunable schemes can also be used for amplitude regeneration of higher-order square QAM modulation format signals. 7.2 Concept of Multilevel Amplitude Regenerator Based on Polarization Wave Mixing The operation principle of our proposed polarization based multilevel amplitude regenerator is shown in Fig. 7 .13. First the noisy star-8QAM signal E" is split into two orthogonal polarization states with a splitting ratio a using a polarization beam splitter (PBS). The splitting ratio is adjusted by a polarization controller (PC) at PBS input. Assuming a < 0.2, at the PBS output, the weaker signal copy ~E .. with polarization state n 2 is almost unaffected while the stronger signal copy .J1 -aE,,, with polarization state n 1 is modified by the self-phase modulation (SPM) effect in a highly nonlinear fiber Multilevel Amplitude Regeneration Diagram Based on Coherent Polarization Mixing Signal Copy Generation ('l i:;· g 5' s ., .,.____ _ Pol·Multiplexing/ Weight Coherent Addition Adjustment Fig. 7 .13. Diagram of polarization-based phase-preserving multilevel amplitude regeneration. 88 (HNLF), which converts the amplitude fluctuations into phase changes cp as given by <p= rLP where Pis the signal power, Lis HNLF length and r is the nonlinearity factor, Note that the two signals are counter-propagating through the HNLF in order to experience the same phase shift and also to reduce the effect of cross-phase modulation (XPM), Two circulators are used to make two different paths for propagating and counter-propagating signals, At the HNLF output the stronger signal electric field E, is (7-1) and the weaker signal electric field Ew is (7-2) Then a custom polarization beam combiner (PBC) applies attenuation x and a phase shift e to the higher power signal, resulting in a considerable vector magnitude difference between the unaffected signal copy and the self-phase modulated copy that is required to get the optimum amplitude regeneration without adding a large component of phase noise. The polarization multiplexed signal Em at the PBC output is (7-3) Since the self-phase modulated polarization state and the unaffected polarization state travel together throughout the system, they remain coherent with each other and a polarizer can be utilized to add the two polarization states coherently [137], This scheme also provides extra tunability to achieve the required magnitude difference by tuning the polarizer angle, The polarizer output E, tuned at angle ¢ with respect to the n 1 axis is 89 E P oc /i E" exp(-j (}I 2) cos(¢)+ Ew sin(¢). (7-4) Amplitude regeneration at each amplitude level is achieved by coherent addition of the SPM-based phase-modulated polarization state and the unaffected one. To achieve significant regeneration strength on both amplitude levels of star-QAM signals, two stages of the proposed module are cascaded to get the desired regeneration results. Noisy Constellation (High Power) I Quadrature SPM-based phase modulated Constellation (Low Power) I Quadrature -----"'~ ~ ·2 A3 A'l I A2 I -----? I Al I Regenerated Constellation I Quadrature ··-..... ~.,>- _1 __ _ I In phase I Fig. 7 .14. Concept of multilevel amplitude regeneration: the original signal is phase modulated based on the self-phase modulation periodic effect and is added coherently to the original signal. Figure. 7 .14 is a conceptual diagram of coherent addition at each amplitude state on a constellation plane in our proposed technique. At the left, we show three vectors corresponding to various amplitude fluctuations located in the noisy constellation point on the in-phase/quadrature (IQ) plane. These three arbitrary vectors A 1 , A 2 and A 3 correspond to the lowest, the medium level and the highest power level of a single arbitrary constellation point on a constellation diagram. A / is the corresponding SPM- based phase rotated copy of Ai and is proportional to Ai exp() 'P;), where CfJ; = yL IAi 1 2 • As 90 can be seen, the A; vector has the largest SPM-based phase rotation, which results in decreasing the A 3 amplitude after coherent polarization mixing, while the A; vector has the smallest SPM-based phase rotation, which results in increasing the A 1 amplitude level after coherent polarization mixing, The result is that the distributed noisy constellation point is mapped into a constellation point with reduced amplitude noise as shown in the right side of Fig, 7,14, 7.2.1 Setup The simulation setup for the multilevel amplitude regeneration scheme is depicted in Fig, 7,15, We use the VPitransmissionMaker and VPicomponentMaker software package from VPiphotonics for this work, An in-phase/quadrature (I/Q) modulator is driven by two electrical data streams to generate a 10- Gbaud quadrature-phase shift keying (QPSK) modulated signal at 1552,5nm with a pseudo-random bit stream period (PRBS) l1 5 -1, followed by an amplitude modulator (AM) to generate a star-8QAM signaL The star- 8QAM signal power ratio is 1:5 and the optical signal-to-noise ratio (OSNR) is adjusted using an amplitude and phase noise emulator, To generate a star-16QAM signal, an I/Q modulator is followed by a phase modulator, The resulting signal is then sent to an erbium doped fiber amplifier (EDFA), polarization controller (PC), and a polarization beam splitter (PBS) to generate two signal copies with orthogonal polarization states, The splitting ratio of PBS is tuned to 85: 15 by adjusting the PC These two polarization states are sent to a HNLF with zero-dispersion wavelength (ZDW) of 1552,5 nm, length of 1000 m and nonlinearity factor of 20 l!W -km in reverse directions, These two copies 91 are multiplexed by a custom PBC after applying attenuation and phase adjustments to the SPM modulated signal copy, then are sent to a polarizer tuned at 45 degrees with respect to the n 1 polarization state for coherent polarization mixing. Note that two regenerator modules are cascaded to get the desired regeneration results for star-QAM signals with two distinct amplitude levels. Star-BQAM Transmitter Data I Data cw Laser DataQ AM .................................................................................. -~ 0 ... a. .0 •> 'i I: :::l (I) Cl (I) - ... ~ C) <.>(I) ;; 't) a. I o~ BPF Custom PBC Polarizer Fig. 7.15. Simulation setup. CW: continuous wave; PC: polarization controller; I/Q Mod: In-phase/quadrature modulator; AM: amplitude modulator; PBS: polarization beam splitter; HNLF: highly-nonlinear fiber; PBC: polarization beam combiner; BPF: band pass filter; EDF A: erbium doped fiber amplifier; VOA: variable optical attenuator. 7.2.2 Multilevel Amplitude Regeneration in a Back-to-Back Schematic: Results and Discussions Figures 7.16 (a), 7.16 (d) and 7.16 (g) depict the constellation diagrams of a noisy star- 8QAM signal with optical signal-to-noise ratio (OSNR) of 20 dB, 22 dB and 24 dB. Figures 7.16 (b), 7.16 (e) and 7.16 (h) show the star-8QAM constellation diagrams after a one-stage amplitude regenerator. As can be seen, the first stage amplitude regenerator is 92 Noisy star-8QAM signal Regenerated signal after one-stage Regenerated signal after two-stage amplitude regenerator amplitude regenerator (a) ·=·t~r~: - (b) ·d~o:, (c) . ............ -~~; :!~ :::~,. High noise >(~:~: :~ ·::~·$1·: -:;.~;: :~~ j '# .: ;~~ ··_ 1: t • l '' -- - (OSNR20 dB) -f~·:. .. ~ . _, ·-· ~ •: . ~'. ·'•~i\ j~ ~i ·.-~f·: .. ~,.., (d) ·r~! ·: (e) -i~· (f) _ ......._ ;*': - ·~ .~· Medium noise ~ - ~~; ~: - -- - If #, f· ·::~.;.~ ".f t l (OSNR22 dB) . . ·:· ~~ - -~v --~·: -~! '"!;Ii . .. ~r -~ (g) ·llt· (h) ·1!~\ (i) x . ··~ - • ._ .,,,.. - - Low noise : tlf.; .~-- " . * ~ : f 1 ; 1' x J 1 (OSNR24 dB _.,, .. ~ ...:;..... :~ ~~~: : ~-· Fig. 7.16. Back-to-back constellation diagrams with OSNR 20 dB: (a) a noisy star-8QAM signal; (b) regenerated signal after one-stage amplitude regenerator; (c) regenerated signal after two-stage amplitude regenerator. Back-to-back constellation diagrams with OSNR 22 dB: (d) a noisy star-8QAM signal; (e) regenerated signal after one-stage amplitude regenerator; (f) regenerated signal after two-stage amplitude regenerator. Back-to-back constellation diagrams with OSNR 24 dB: (g) a noisy star-8QAM signal; (h) regenerated signal after one stage amplitude regenerator; (i) regenerated signal after two-stage amplitude regenerator. - -~~ J x Jx intentionally tuned to apply the maximum amplitude n01se reduction on the lower- amplitude level. Figures 7 .16 ( c ), 7 .16 (f) and 7 .16 (i) show the star-8QAM constellation diagrams after a two-stage amplitude regenerator in which the second stage is tuned to apply the maximum amplitude noise reduction on the higher-amplitude level. The performance of the system is evaluated by measuring regeneration factor versus OSNR for both amplitude levels of a star-8QAM. The regeneration factor is defined as the ratio of input amplitude noise standard deviation (STD) to output amplitude noise standard deviation. 93 6 ... 5.5 0 5 - ~ 4.5 LL c: 4 .2 3.5 «) ... 3 Q) c: 2.5 ~ Q) 2 a: 1.5 1 ,-• ... , ..... ,A' ...... . .. ... .. .- ....... - • - lower-amplitude level (star-SQAM) - • - higher-amplitude level (star-SQAM) .. ; .... , ..... ..... .... ~--.. 16 17 18 19 20 21 22 23 24 25 Input SignaJ OSNR (dB) Fig. 7.17. Regeneration factor vs. OSNR for both amplitude levels of a regenerated star- 8QAM signal after a two-stage amplitude regenerator in a back-to-back configuration. As shown in Fig. 7.17, a peak of almost 5 at 20 dB OSNR for the higher-amplitude level of star-8QAM and a peak of almost 3 at 22 dB OSNR for the lower-amplitude are achieved. The regeneration factor is more than 3 .5 for the higher-amplitude level of star- 8QAM at 17 dB input OSNR. This shows the regeneration strength of the proposed multilevel amplitude regeneration module in the wide range of input signal OSNR. Figure 7.18 shows the power transfer function of the one-stage regenerator. It is measured by applying a noisy 8-PSK signal as the input of the proposed amplitude regenerator. The power transfer function has two plateau regions with the input power difference of almost 4.7 dB. Note that we cascade two of the proposed polarization-based regenerators to get the maximum regeneration factor for both amplitude levels. 94 20 ...... - f m ' 1J 18 ! - ' ... Q) • 3:: I 0 16 J a.. .. ....... ..-" - :::J a.. ~· - 14 :::J • 0 ~ , , 12 18 20 22 24 26 28 30 Input Power (dB) Fig. 7 .18. Power transfer function of the one-stage polarization-based amplitude regenerator. The phase-preserving multilevel amplitude regeneration concept 1s evaluated by implementing the scheme on 10 Gbaud noisy star-16QAM signals as well. Figures 7.19 (a) and 7.19 (d) depict the constellation diagrams of a noisy star-16QAM signal with optical signal-to-noise ratio (OSNR) of 22 dB and 24 dB and OSNR bandwidth of 6.25 GHz. Figures 7.19 (b) and 7.19 (e) show the star-16QAM constellation diagrams after a one-stage regenerator. Figures 7.19 (c) and 7.19 (f) show the star-16QAM constellation diagrams after a two-stage polarization-based multilevel amplitude regenerator in which the first stage amplitude regenerator is intentionally tuned to apply the maximum amplitude noise reduction on the lower-amplitude level while the second stage is tuned to apply the maximum amplitude noise reduction on the higher-amplitude level. As can be seen, amplitude regeneration is achieved at both levels without noticeable phase noise. 95 Noisy star-1 GQAM signal After one-stage amplitude regenerator After two-stage amplitude regenerator (a) -'~); ·1k '~ .:.•.ill_; _ _ ~, _-_ •. • _ _ >~~:/~~ :.~," "2l -~~<-. _.;~ (c) x .··~ . :tjfi-~':':j~ ~j ,: ~'*; (OSNR22dB) ~~- :- .~- -- : "··· ~ilif:: '¥ -: it" *; · *; -~' .. · .· . -~. - -·_-~ , - '4fi! ~ l~ "rf'I '1f, (e) .. \!!~~: ·-.£~1f.~· :H.?it ... (~ ~-': • x~ (OSNR24dB) .~ Fig. 7.19. Back-to-back constellation diagrams with OSNR 22 dB: (a) a noisy star- 16QAM signal; (b) regenerated signal after one-stage amplitude regenerator; (c) regenerated signal after two-stage amplitude regenerator. Back-to-back constellation diagrams with OSNR 24 dB: (d) a noisy star-16QAM signal; (e) regenerated signal after one-stage amplitude regenerator; (f) regenerated signal after two-stage amplitude regenerator. The demonstrated results in this chapter is based on the fact that two regenerator modules are cascaded to get the desired regeneration results for star-QAM signals with two distinct amplitude levels. However, even one stage of the proposed multilevel amplitude regeneration module can be used to apply fair level of regeneration to both amplitude levels. 96 7.2.3 Multilevel Amplitude Regeneration in a Nonlinear Transmission Line Schematic Results and Discussions The impact of having a multilevel amplitude regenerator before a transmission line to avoid nonlinear phase noise originating from the Gordon- Mollenauer effect is shown by implementing the scheme on 10 Gbaud noisy star-8QAM and star-16QAM signal with With pre-regeneration x~, ·If ,• .-.~. ! , f (, N a: z en 0 . J · .. , --~ ' . . ·' .. "·" :"f< - (a With pre-regeneration x ··'·~. - ctl f ~ x .. -~ cc: z CJ) 0 - -~· -·~ - ~ . .._ ... x (c) - ctl 'O N N cc: z CJ) 0 - No p_re~regeneration ... : . ; . ..§~ ~i~*·:;·:·;·~~~:·~.~ . ·.~ : ' ~ .. ; ~ . /, . (b No pre-regeneration .. ...;;;~l(::'Ji·;,~;,., .. , •\ . . .. i-<'f·~·. ' ' . ·.· !, ' : : ' ... .... . ~\;· i/ t ''·\. l' . -~,. .· :;.,.~. ., '· l~· ·;~i " i· ~. i ·,~ ;/ ' ' · l ·t · ··. , ,,, . ·~~f . :~ : ·: ·1 y;f) .. ,( f1' : > ... · r ' !~ .... • ; : '• ~ . 'I . '. h, · . . , • ' . '~ ~ (\~,r~~"'j.',(' :· (d) Fig. 7.20. Transmitted constellation diagrams of star-8QAM after 150 km transmission line with/without a pre-regeneration module at the input of a nonlinear transmission line. 97 almost 16 dBm average power transmitted through a 150 km nonlinear transmission line. A pre-amplifier and the transmission line are inserted between the polarizer and EDF A in Fig. 7.15. Figures 7.20 (a) and 7.20 (c) depict the constellation diagrams of noisy star- 8QAM signals with signal-to-noise ratio (OSNR) of 20 dB and 22 dB after a 150 km transmission line with nonlinear index of 2.6e - 20 m 2 /W and core area of 80e -12 m 2 with a two-stage proposed amplitude regenerator module (pre-regeneration) located at the transmission line input terminal. With pre-regeneration - Cl "C o:::t" N a: z CJ) 0 ._... f x·: \ .. x · x . # (c) Fig. 7.21. Transmitted constellation diagrams of star-16QAM after 150 km transmission line with/without a pre-regeneration module at the input of a nonlinear transmission line. 98 Figures 7.20 (b) and 7.20 (d) show the transmitted star-8QAM constellation diagrams having the same noise characteristic with no amplitude regenerator module (no pre-regeneration) at the transmission line input terminal. Figures 7.21 (a) and 7.21 (c) also depict the constellation diagrams of noisy star-16QAM signals with optical signal- to-noise ratio (OSNR) of 22 dB and 24 dB after a 150 km transmission line with a two- stage multilevel amplitude regenerator located at the transmission line input terminal, while Figs. 7.21 (b) and 7.21 (d) show the transmitted star-16QAM constellation diagrams having the same noise characteristic with no amplitude regenerator module (no pre-regeneration) at the transmission line input terminal. As can be seen, when the amplitude regenerator module is removed from the transmission line input terminal; considerable amount of nonlinear phase noise generated due to the Gordon- Mollenauer effect. The demonstrated results show the critical role of having amplitude regeneration modules at the transmission line input terminal. 6 5 ... 0 - 0 4 cu u. Q) 3 "' ·o z 2 Q) "' cu 1 ..c a.. 0 ·-- ---. ... ... ..... ' ... . ... ... .... 11:·-·-·- &.. -----.~ 4-· --· --· ----- · -·- -- ----. -- ........ · -~ · - -- -- -- 17 18 19 20 21 22 23 24 25 Input Signal OSNR (dB) - • - lower-amplitude level with no pre-regeneration - • - higher-amplitude level with no pre-regeneration - - lower-amplitude level with pre-regeneration - • - higher-amplitude level with pre-regeneration Fig. 7.22. Phase noise factor vs. input signal OSNR for both amplitude levels of a transmitted star-8QAM with/without pre-regeneration module at the input of a nonlinear transmission line. 99 To show the level of nonlinear phase noise generated by the Gordon-Mollenauer effect in a transmission line, we define the phase noise factor. The phase noise factor is defined as the ratio of transmitted signal phase noise standard deviation to initial signal phase noise standard deviation. As shown in Fig. 7.22, the phase noise factor for the higher- amplitude level of star-8QAM with no pre-regeneration has a peak value of almost 5.5, while the peak value reduces to almost 2 by applying pre-regeneration. The phase noise factors are almost in the same order for lower-amplitude with no pre-regeneration and higher-amplitude with pre-regeneration. Figure 7 .23 shows the polarizer angle vs. attenuation value embedded inside the PBC at BER 1 Oe-4. As can be seen, by applying 30 dB attenuation and tuning the polarizer angle to 45 degrees; or by applying almost 17 dB attenuation and tuning the polarizer angle to almost 13 degrees at the first stage of 50 Q) 40 (b) C> c <( 30 c 0 - 20 m N ... m 0 10 a.. 0 0 10 20 Attenuation( dB) regenerator _._first stage of amplitude regenerator 30 Fig. 7.23. Polarizer angle vs. attenuation value embedded in a custom polarization beam combiner (PBC) at BER 1 Oe-4. polarization-based regenerator, the same BER of 1 Oe-4 can be achieved, thus showing additional degrees of freedom for tunability in our proposed regenerator. 100 Figures 7.24 (a) depicts the constellation diagrams of a noisy square-16QAM signal with optical signal-to-noise ratio (OSNR) of 22 dB. Figures 7.24 (b) shows the square-16QAM constellation diagrams after a two-stage amplitude regenerator in which Noisy standard-16-QAM signal .... ~ · :· . . -..: ·:~ . .. , . :· . · · ~~: : . . ... , . . ;;~;: · ·. \{;f. '·· · ·~: . ;~;)F: : · ,• ... After two-stage amplitude regenerator I>/ .Jfr · .. . :',· .,~;.· ~ f. . :~,k . .. ~ ... ·. 1'· ;·:~ .. .. . ::~: \ ?~ :·, •. ~ '.- ·~ . . •:'·. * ~,, ... 7: x: · : '· -~- ~.~·· . .,, '' (b) Fig. 7.24. Back-to-back constellation diagrams with OSNR 22 dB: (a) a noisy square-16QAM signal; (b) regenerated signal after two-stage amplitude regenerator. the first stage of amplitude regenerator is intentionally tuned to apply the maximum amplitude noise reduction on the mid-amplitude level while the second stage is tuned to apply the maximum amplitude noise reduction on the higher-amplitude level. As can be seen the amplitude noise removed from the mid-amplitude level and the higher-amplitude level while the lower-amplitude level remains almost unaffected. As it is discussed before amplitude noise removal from the higher amplitude levels is a priority as they get affected more by the Gordon- Mollenauer effect in a nonlinear transmission line. 101 7.3 Dispersion Tolerance Enhancement Using Optical Duobinary Detection in an Optimized 20-70 Gbit/s NRZ-OOK Transmission The demand for high capacity optical communication services has led to the increase in data speeds of transmitter and receiver terminals. The increase of line rate up to 40-Gbit/s fundamentally decreases chromatic dispersion tolerance of a system by a factor of 16 compared to a 10-Gbit/s system under the same conditions [103]. The power spectral density of a signal can be narrowed at the transmitter or receiver side by using an electrical low-pass filter with a 3-dB cut off frequency of approximately 0.25 of the bit rate. This is known as the duobinary modulation format. This technique has been proposed as a solution to improve the chromatic dispersion tolerance of a system [104- 107]. Typical methods for generating duobinary data include: (a) sending electrical NRZ data through an electrical low-pass filter with a 3-dB cut off frequency of approximately 0.25 of the bit rate at the transmitter side [104]; (b) sending optical NRZ data to an electrical low pass filter with a bandwidth of approximately 0.25 of the bit rate just before detection at the receiver side [105]; (c) sending optical NRZ data to a delay line interferometer (DLI) just before detection at the receiver side [105]. In this section, we demonstrate optical duobinary detection for an optimized NRZ-OOK transmission. We show that for an optimized electrical NRZ voltage and a proper modulator bias voltage at transmitter side; optical 3-level intensity modulation format can be detected. This results from the low pass profile of chromatic dispersion effect in optical fiber. This results in having a longer maximum reach, and can be considered as a low cost way of dispersion tolerance enhancement. Fully optical three-level intensity detection with proper data pre- coding might have applications in optical networks. 102 Fully-Optical NRZ to 3-Level Amplitude (Duobinary) Signal Conversion Optimized NRZ transmitter M,ld Zeh 1dl'r r JlaL Low Pass Dispersion Profile b ... F requency H ( f) = Cos ( 7r LD A 6 f 2 I c) Duobinary Signal Fig.7.25 (a) Optical duobinary detection scheme. Electrical NRZ Signal L Conventional Duobinary Signal Generation E lectrical Delay Line Interferometer (DLI) 3-Level Amplitude (Duobinary) Signal I bit delay (T) O+O 0+1 1+ 1 Electrical Low Pass Filter .............. BW:bit rate/2 .......,,. Fig. 7 .25 (b) Conventional electrical duo binary detection scheme using a DLI. Figure 7.25 (a) shows a conceptual diagram of the proposed dispersion tolerance enhancement technique using fully-optical duobinary detection based on an optimized operation of an optical modulator (MZM) at the transmitter side in the linear region of 103 transfer function. Here vbuu = v""";caz = v/ 4, vrr = 4volt. The fiber chromatic dispersion frequency response is given by where So we have 8 2 k - )} D(J) aa} 27fc H(f) = cos(m'i/T) T = DzJ24f c (7-5) (7-6) (7-7) (7-8) where D(J) is a dispersion parameter, and c is the speed of light. Figure 7.25 (b) shows a conceptual diagram of a conventional electrical method for dispersion tolerance enhancement using a delay line interferometer (DLI). It can be shown that the frequency response of a DLI element is [105] cos(efF) (7-9) where T represents the DLI one bit delay. As can be seen there is a duality between the DLI configuration and fiber chromatic dispersion profile. For each bit rate, there is a minimum and maximum length at which complete NRZ-OOK to 3-level intensity modulation format conversion occurs. This method might be considered as a low cost fully-optical dispersion tolerance enhancement technique for short and midrange transmission. 104 7.3.1. Results and Discussions To demonstrate the concept, we use a continuous-wave (CW) laser at 1555.3 nm modulated using a Mach-Zehnder modulator (MZM) to generate NRZ-OOK channel at bit rate up to 70-Gbit/s, with the modulator operating in the linear region. The generated OOK is then sent to a spool of standard single mode fiber (SMF), with a zero-dispersion wavelength of -1312 nm, and the optical eye diagrams is captured at the output for different lengths and bit rates. 40Gb/s Standard 40Gb/s Standard 40G bis B-B 20G bis after 30Gbls after NRZ rr NRZ after 10km KK -<1km 15km H~ 40Gb/s Optimized 40Gb/s Optimized 40G bis after 50G bis after 70G bis after tt~ ae& (a) (b) Fig. 7.26. (a) Standard and optimized optical eye diagrams of 40Gbit/s NRZ-OOK at lOkm; (b) Optical eye diagrams of an optimized NRZ-OOK to 3-level intensity modulation (duobinary) format conversion for various bit rates at 1555.3 nm. Figure 7.26(a) shows optical eye diagrams of standard and optimized 40Gbit/s NRZ-OOK after IOkm. As can be seen the optical eye diagram of a standard NRZ-OOK is completely distorted after IOkm, while the conversion to 3-level intensity modulation 105 'E 1500 c: <n .3. c: 1000 0 -~ Q) a... 500 en i:5 0 D ispersion Map 0 1 0 2 0 30 4 0 50 60 7 0 80 D is ta n c e [km] Channe ls A.= 1 .5303 ~·m A.=1 .5553 ~·m A.= 1 .6 2 0 3 p m Fig.7.27. Dispersion vs. distance at various wavelengths in a 20 Gbit/s system. format occurs after IOkm for an optimized NRZ-OOK signal. Figure 7.26(b) shows the optimum eye diagrams for NRZ-OOK to duobinary format conversion for various bit rates. 80 70 -;:;-60 QJ ~ so :a Q.40 QJ 'l;j 30 0:: ~ 20 10 0 l ,, ... ' I \ r------.·~ ' _ .... -- ....... ~ .... -~ ..... - 0 20 40 Length( km) 60 - - M aximum Reach( km) - - Minimum Length( km) Fig.7.28. Maximum and minimum possible lengths vs. bit rate for full NRZ-OOK to 3- level intensity modulation format conversion; (insets) Optical eye diagrams of 40 Gbit/s duo binary signals at 8km, 11 km and 7km. As can be seen in Fig 7.27, to have a full format conversion in a 20Gbit/s system, almost 700 ps/nm dispersion is needed, which is equivalent to 41 km at 1555.3 nm , 44 km at 1530.3 nm and 34 km at 1620.3 nm. Note that in our simulation the MZM transfer 106 funct ion is considered as sin 2 [ n12(v _.,.,,+v • ...,) 1v J so that the modulated signal w ill have the same polarity as the original binary sequence. Fiwre 7.28 shows the maximum and minimu m possible lengths versus bit rate at which full NRZ-OOK to duo binary format conversion occurs. As shown in the insets, after 7km the conversion is not yet comp lete and the minimum length for full conversion at 40Gb/ s is 8k m, while the duob inary sigpal can be detected until the maximum length of 1 l km. =- 3.5 !. 3 ., "' " 2.5 ... l 2 ~ 1.5 5 1 'S 0.5 i 0 0 - • . ... . . . s 10 Maximum i:each (Ian) . 1S • Blas: IV • Blas: 2V A Blas: 3 V Fig.7.29. Electrical NRZ amplitude (volt) vs. maximum reach for NRZ-OOK to duobinary format conversion for various bias voltages in 40 GB/s transmission system with vn = 4. Fiwre 7.29 shows the impact of e lectrical NRZ a mplitude on maximum possible reach of optical duobinary detection for various bias vohages in 40Gbit/s transmission syste m with vn = 4. Here the operation is in linear region and should not pass the null point. As can be seen, the maximum reach of 11 km can be achieved by applying 1 volt driving voltage at bias points of 1 voh and 2volt. 107 System bitrate Minimum length Maximum reach Margin 20Gb/s 34 km 45 km 11 km 30Gb/s 15 km 20 km 5 km 40Gb/s 8 km 11 km 3 km 50Gb/s 7km 9km 2 km 60Gb/s 4 km 5km 1 km 70Gb/s 3 km 3.5 km 0.5 km Table. 7.1. Margin between the minimum length and maximum reach possible for optimum format conversion. The margin between the minimum length and maximum reach possible for the optimum NRZ-OOK to duobinary format conversion is shown in Table.7.1. 108 7.4 Concept of Phase-Transparent NOLM-Free Multilevel Amplitude Regeneration Based on Polarization Wave Mixing The operation principle of our proposed polarization-based NOLM-free multilevel amplitude regenerator is shown in Fig. 7.30. First the noisy star-8QAM signal E in is split into two orthogonal polarization states with a splitting ratio a using a polarization beam splitter (PBS). The splitting ratio is adjusted by a polarization controller (PC) at PBS input. Assuming a < 0.2 , at the PBS output, the weaker signal copy .f-;;Ein with polarization state n 2 is almost unaffected while the stronger signal copy ,J1 - a E ,,, with polarization state n 1 is modified by the self-phase modulation (SPM) effect in a highly nonlinear fiber (HNLF), which converts the amplitude fluctuations into phase changes q; as given by <p= yLP where P is the signal power, L is HNLF length and r is the nonlinearity factor. .----------------------------------------------------------------------------------------------------------------- Noisy Constellation q r----@--1 fii j Phaseshifler I __ , I I Beam Splitter I l---- Polarizatio Beam Combiner -·· ··· I r-- - tt I I A I H NLF I n, ~-<ID.J Regenerated Constellation q Polarizer '-------------------------------------------------------------------------------------------------_j Fig. 7.30. Diagram of NOLM-free polarization-based phase-preserving multilevel amplitude regeneration. 109 Note that just the stronger signal is propagating through the HNLF. In order to experience the same propagation delay the required amount of phase shift is applied to the weaker signal. At the HNLF output the stronger signal electric field E. is E, =.JI-a Em exp(-) yL(l-a)IEml 2 ) (7-10) and the weaker signal electric field Ew after the phase shifter is (7-11) Then a custom polarization beam combiner (PBC) applies attenuation x to the stronger signal, resulting in a considerable vector magnitude difference between the unaffected signal copy and the self-phase modulated copy that is required to get the optimum amplitude regeneration without adding a large component of phase noise. The polarization multiplexed signal Em at the PBC output is (7-12) This scheme also provides extra tunability to achieve the required magnitude difference by tuning the polarizer angle. The polarizer output E, tuned at angle ¢ with respect to the ii 1 axis is EP oc Ji E, cos(¢)+ Ew sin(¢). (7-13) Amplitude regeneration at each amplitude level is achieved by coherent addition of the SPM-based phase-modulated polarization state and the unaffected one. To achieve significant regeneration strength on both amplitude levels of star-QAM signals, two stages of the proposed module are cascaded to get the desired regeneration results. 110 7.4.1 Setup The simulation setup for the second multilevel amplitude regeneration scheme is depicted in Fig. 7.31. We use the VPitransmissionMaker and VPicomponentMaker software package from VPiphotonics for this work. An in-phase/quadrature (I/Q) modulator is driven by two electrical data streams to generate a 10-Gbaud quadrature-phase shift keying (QPSK) modulated signal at 1552.5nm with a pseudo random bit stream period (PRBS) 2 15 -1, followed by an amplitude modulator (AM) to generate a star-8QAM signal. The star-8QAM signal power ratio is 1:5 and the optical signal-to-noise ratio (OSNR) is adjusted using an amplitude and phase noise emulator. The resulting signal is then sent to an erbium doped fiber amplifier (EDFA), polarization controller (PC), and a polarization beam splitter (PBS) to generate two signal copies with orthogonal polarization states. The splitting ratio of PBS is tuned to 85:15 by adjusting the PC. The higher power polarization state is sent to a HNLF with zero-dispersion wavelength (ZDW) of 1552.5 nm, length of 1000 m and nonlinearity factor of 20 l!W - km . The required phase shift is applied to the lower power polarization state to remain coherent with the other polarization state. Note that the phase shifter is tuned at almost 320 degrees to apply the required phase shift to the unaffected copy of the signal. These two copies are multiplexed by a custom PBC after applying attenuation to the SPM modulated signal copy, then are sent to a polarizer tuned at 45 degrees with respect to the n 1 polarization state for coherent polarization mixing. Note that two regenerator modules are cascaded to get the desired regeneration results for star-QAM signals with two distinct amplitude levels. 111 Star-BOAM Transmitter cw Laser Custom PBC Fig. 7.31. Simulation setup. CW: continuous wave; PC: polarization controller; I/Q Mod: In-phase/quadrature modulator; AM: amplitude modulator; PBS: polarization beam splitter; HNLF: highly-nonlinear fiber; PBC: polarization beam combiner; BPF: band pass filter; EDFA: erbium doped fiber amplifier; PS: phase shifter; VOA: variable optical attenuator. 7.4.2 Multilevel amplitude regeneration in a back-to-back schematic Results and Discussions The phase-preserving multilevel amplitude regeneration concept 1s evaluated by implementing the second scheme on 10 Gbaud noisy star-8QAM signals as well. Figures 7.32 (a), 7.32 (d) and 7.32 (g) depict the constellation diagrams of a noisy star-8QAM signal with optical signal-to-noise ratio (OSNR) of 20 dB, 22 dB and 24 dB. Figures 7.32 (b), 7.32 (e) and 7.32 (h) show the star-8QAM constellation diagrams after a one-stage amplitude regenerator. As can be seen, the first stage amplitude regenerator is intentionally tuned to apply the maximum amplitude noise reduction on the lower- amplitude level. Figures 7.32 (c), 7.32 (f) and 7.32 (i) show the star-8QAM constellation diagrams after a two-stage amplitude regenerator in which the second stage is tuned to apply the maximum amplitude noise reduction on the higher-amplitude level. 112 Noisystar-BQAM signal (a) ·'.~- ...... ::'4i ~i · ;~{ :~·:.':ft': al "O 0 N a: z (/) Q. (d) :·•· .;.~ ·~ ~~· ,...... al "O N N a: z Q, (g) • m "O "'" N a: z (/) Q, . * ' :~ --· ··~·~·. ... ;. "f ~: .Jlt- ·~·· ·; .. ; * ·;5 :· · · ·'. ... - .. ': .... ~ -:'- f~ . • . ~ · .: .. :.i. · · - · : (#.: Regenerated signal after stage one (b) . ·~~" :';~ . "'$ j ·· ~.r · '~, : , ~: .. \., ;::: .. :~2 ; · '- · ri" "O 0 N a: z (/) Q. (e) . ... . " . •• •• ,' :!· .. ''' '', . . · '-· ~ .. ~11:~: . .. i·"'. ~ o i •,'.f' I m " ~ .. . ~ . . N a: z (/) Q. (h) i ........ :- -= .. al "O "'" N a: z (/) Q, •• " "-•" ·'7!'· '·~ ,4 · • .·· . · ii _.i " ·~ ·· w : . ,. Regenerated signal after stage two (c) ' ;· : j; ........ ·;: ,$: :< al " "O 0 N a: z (/) Q. (f) (i) i ,,. ~ ,...... al "O "'" N a: z (/) Q. .::~;· ·~•· .''' ".· ~ .. : ,;]IL!' " ·-~· : •~: .. .... . .... ~ • • • • .... ,. . • l ~ t Fig. 7 .32. Back-to-back constellation diagrams with OSNR 20 dB: (a) a noisy star-8QAM signal; (b) regenerated signal after one-stage amplitude regenerator; (c) regenerated signal after two-stage amplitude regenerator. Back-to-back constellation diagrams with OSNR 22 dB: (d) a noisy star-8QAM signal; (e) regenerated signal after one-stage amplitude regenerator; (f) regenerated signal after two-stage amplitude regenerator. Back-to-back constellation diagrams with OSNR 24 dB: (g) a noisy star-SQAM signal; (h) regenerated signal after one stage amplitude regenerator; (i) regenerated signal after two-stage amplitude regenerator. 113 Note that the phase shifter is tuned at almost 320 degrees to apply the required phase shift to the unaffected copy of the signal. The performance of the system is evaluated by measuring regeneration factor versus OSNR for both amplitude levels of a star-8QAM. Again the regeneration factor is defined as the ratio of input amplitude noise standard deviation (STD) to output amplitude noise standard deviation. 6 5.5 ... 5 0 ti C1I LL 4.5 c 4 0 3.5 ·:;; Qi 3 c 2.5 Cll Cl Cll 2 a: 1.5 1 16 • , .... A' ..... - · - lower-amplitude level (star-8QAM) ....... higher-amplitude level (star-8QAM) . .. .. .. . -- .... ... ::: . , . . . . .... - ·- 17 18 19 20 21 22 23 24 25 Input Signal OSNR (dB) Fig. 7.33. Regeneration factor vs. OSNR for both amplitude levels of a regenerated star-8QAM signal after a two-stage amplitude regenerator in a back-to-back configuration. As shown in Fig. 7.33, a peak of almost 3.5 at 20 dB OSNR for the higher-amplitude level of star-8QAM and a peak of almost 2.3 at 22 dB OSNR for the lower-amplitude are achieved. The regeneration factor is almost 2.5 for the higher-amplitude level of star- 8QAM at 17 dB input OSNR. This shows the regeneration strength of the proposed multilevel amplitude regeneration module in the wide range of input signal OSNR. The demonstrated results in this chapter is based on the fact that two regenerator modules are cascaded to get the desired regeneration results for star-QAM signals with two distinct amplitude levels. However, even one stage of the proposed multilevel amplitude 114 regeneration module can be used to apply fair level of regeneration to both amplitude levels. 7.4.3 Multilevel Amplitude Regeneration in a Nonlinear Transmission Line Schematic Results and Discussions The impact of having a multilevel amplitude regenerator before a nonlinear transmission line to avoid nonlinear phase noise originating from the Gordon-Mollenauer effect is With pre-regeneration . ,. iii"" ~- - "O C'\I C'\I cc z CJ) Q. No pre-regeneration ~:·= · -~ .. ::".C·.~:~ ·~~~~~.•r,_:: ,.~. . : ~:~~· ··.;;~·:· ·:&_ :-.· . · .. ·· ~· i~ . • • (b) Fig. 7.34. Transmitted constellation diagrams of star-8QAM after 150 km transmission line with/without a pre-regeneration module at the input of a nonlinear transmission line. 115 shown by implementing the second proposed regeneration scheme on 10 Gbaud noisy star-8QAM signals with almost 16 dBm average power transmitted through a 150 km transmission line. A pre-amplifier and the transmission line are inserted between the polarizer and EDFA in Fig. 31. Figures 7.32 (a). 7.32 (c) and 7.32 (e) depict the constellation diagrams of noisy star-8QAM signals with signal-to-noise ratio (OSNR) of 20 dB. 22 dB and 24 dB after a 150 km transmission line with nonlinear index of 2.6e - 20 m 2 IW and core area of 80e -12 m 2 • with a two-stage amplitude regenerator (pre-regeneration) located at the transmission line input terminal. Figures 7.32 (b). 7.32 (d) and 7.32 (f) show the transmitted star-8QAM constellation diagrams having the same noise characteristic with no amplitude regenerator module (no pre-regeneration) at the transmission line input terminal. As can be seen. when the amplitude regenerator module is removed from the transmission line input terminal; considerable amount of nonlinear phase noise generated due to the Gordon-Mollenauer effect. The demonstrated results show the critical role of having an amplitude regeneration module at the transmission line input terminal. 116 Chapter 8: Conclusion With growing demand for higher transmission capacity, along with the desire to lower the cost per information bit, spectrally-efficient advanced modulation formats have become popular. Wavelength-division multiplexing (WDM) systems along with optical quadrature amplitude modulation (QAM) formats seem to be the promising methods for data transmission in long-haul optical transmission lines. For that reason optical signal processing modules that are compatible with advanced modulation formats are expected to be common and useful in future optical networks. In this dissertation, we introduced and experimentally demonstrated two methods for generation of QAM signals. In the first method that we call "polarization-based QAM generation technique," the initial channels are at the same frequency and the amplitude control is implemented outside the integrated device using a polarizer. This polarization-based technique offers a wide range of tunability compared to previous proposed methods. It is potentially capable of generating different QAM formats such as 16-QAM and 64-QAM using the same module. We showed that 16-QAM signals at up to 160 Gbit/s could be generated using this technique. In the second method, the initial channels come from different sources with different frequencies. Using this nonlinearity- based technique, we showed multiplexing of four 20Gbit/s OOK channels into a 117 80 Gbit/s 16-QAM. We also showed multiplexing of three 1 OGbit/s OOK channels into a 30 Gbit/s diagonal and rectangular 8-level phase/amplitude (8-PAM) signal. In this dissertation. we introduced and demonstrated all-optical methods to apply various signal processing operations on QAM signals such as demultiplexing and information extraction. These all-optical signal processing modules potentially eliminate the need for bulky. inefficient and expensive coherent receivers in the middle nodes in future optical networks. We experimentally showed symbol information extraction from a 40 Gbit/s 16-QAM signal in the form of a 3-level amplitude modulated signal implementing logic/arithmetic relations between symbols carried by a 16-QAM signal. We also proposed a method for demultiplexing of a QPSK signal into two BPSK sub channels based on phase-sensitive amplification technique. A disadvantage of higher-order QAM signals is higher sensitivity to nmse accumulation. Amplitude noise not only reduces signal quality but may also be converted into nonlinear phase noise in a transmission line due to the Gordon-Mollenauer effect. Cross-phase modulation in wavelength-division multiplexing (WDM) systems is another cause of amplitude noise to nonlinear phase noise conversion as well. This noise is not easy to remove due to the current phase regeneration scheme limitations. In this dissertation. we introduced and developed two schemes for amplitude regeneration of multilevel modulation formats such as star-8QAM. star-16QAM and square 16-QAM. These polarization-based techniques offer a wide range of tunability and simpler schemes compared to previous methods. We obtained the regeneration factor peak of almost 5 at 20 dB OSNR for the higher-amplitude level of star-8QAM signals. and a peak of almost 3 at 22 dB OSNR for the lower-amplitude level. By placing our proposed multilevel 118 amplitude regeneration module at the input of a 150 km transmission line, we showed that the generated nonlinear phase noise level for the higher-amplitude level of star-8QAM is reduced by a factor of almost 2,5, These results demonstrate the significant role that multilevel amplitude regenerators may have in future optical networks, 119 References [1] G. P. Agrawal. Fiber-Optic Communication Systems 3rd ed. New York: Wiley Interscience. 2002. [2] J. M. Kahn and K.-P. Ho, "Spectral efficiency limits in DWDM systems," in 31 8 ' European Conference on Optical Communication (ECOC 2005), Glasgow, vol. 4,pp. 843 - 846, September 2005. [3] K. Petermann and S. Randel, "Strategies for spectrally efficient optical fiber communication systems with direct detection," in International Conference on Transparent Optical Networks (ICTON 2003), Warsaw, Poland, vol. 2, pp. 58 - 63, 2003. [4] H. Louchet, K. Petermann, A. Robinson, and R. Epworth, "On the spectral information distribution in optical fibers," in IEEE Lasers and Electro-Optics Society Annual Meeting (LEOS 2004), Rio Grande, Puerto Rico, vol. 1, pp. 17 - 18, 2004. [5] N. Kikuchi, "Amplitude and phase modulated 8-ary and 16-ary multilevel signaling technologies for high-speed optical fiber communication," in SPIE Proceedings, vol. 6021, pp. 17 - 18, 2005. [6] S. K. Ibrahim, S. Bhandare, and R. Noe, "Performance of 20 Gbit/s Quaternary Intensity Modulation Based on Binary or Duobinary Modulation in Two Quadratures With Unequal Amplitudes," IEEE Journal of Selected Topics in Quantum Electron ics, vol. 12, no. 4, pp. 596 - 602, 2006. 120 [7] R. Noe, "PLL-free synchronous QPSK polarization multiplex/diversity receiver concept with digital I&Q baseband processing," IEEE Photon. Technol. Lett., vol. 17, no. 4,pp.887-889,2005. [8] E. Ip and J. M. Kahn, "Feed forward earner recovery for coherent optical communications," J. Llghtw. Technol., vol.25, no.9, pp.2675-2692, 2007. [9] L. G. Kazovsky, "Performance analysis and laser linewidth requirements for optical PSK heterodyne communications systems," J. Lightw. Technol., vol. 4, no. 4, pp. 415- 425, 1986. [10] C. Wree, D. Becker, D. Mohr, and A. Joshi, "Coherent receivers for phase-shift keyed transmission," in Proc. OFC/NFOEC, 2007, paper OMP6. [11] K. Kikuchi, T. Okoshi, M. Nagamatsu, and N. Henmi, "Degradation of bit-error rate in coherent optical communications due to spectrum spread of the transmitter and the local oscillator," J. Lightw. Technol., vol. LT-2, no.6, pp.1024-1033, 1984. [12] S. Norimatsu and K. Iwashita, "Damping factor influence on linewidth requirements for optical PSK coherent detection systems," J. Lightw. Technol., vol. 11, no. 7, pp. 1266-1233, 1993. [13] C. Xu, X. Liu, and X. Wei, "Differential phase-shift keying for high spectral efficiency optical transmissions," IEEE J. Se!. Topics in Quantum Electron, vol.10, no.2, pp.281-293, 2004. [14] I. Dedic, "56 GS/s ADC: Enabling lOOGbE," in Proc. OFC/NFOEC, 2010, paper OThT6. 121 [15] P. Schvan. J. Bach. C. Fait. P. Flemke. R. Gibbins. Y. Greshishchev. N. Ben Hamida. D. Pollex. J. Sitch. S.-C. Wang. and J. Wolczanski. "A 24GS/s 6bADC in 90nm CMOS," in Proc. IEEE Int. Solid-State Circuits Conf., 2008, pp.544--034. [16] P. J. Winzer, "High-spectral-efficiency optical modulation formats," J. Lightwave Technol.30(24), 3824-3835 (2012). [17] A. Sano, H. Masuda, T. Kobayashi, M. Fujiwara, K. Horikoshi, E. Yoshida, Y. Miyamoto, M. Matsui, M. Mizoguchi, H. Yamazaki, Y. Sakamaki, and H. Ishii, "69.1- Tb/s (432x171-Gb/s) C- and Extended L-Band transmission over 240 km using PDM-16- QAM modulation and digital coherent detection," in Proc. OFC/NFOEC, 2010, paper PDPB7. [18] X. Zhou, J. Yu, D. Qian, T. Wang, G. Zhang, and P. D. Magill, "High-spectral efficiency 114-Gb/s transmission using Po1Mux-RZ-8PSK modulation format and single ended digital coherent detection technique," J. Lightw. Technol., vol. 27, no. 3, pp. 146- 152, 2009. [19] K.-P. Ho, Phase-Modulated Optical Communication Systems. New York: Springer, 2005. [20] M. Seimetz and C.-M. Weinert, "Options, feasibility, and availability of 2x4 90 hybrids for coherent optical systems," J. Lightw. Technol., vol. 24, no. 3, pp. 1317- 1322, 2006. [21] Y. Han and G. Li, "Coherent optical communication using polarization multiple input- multiple-output," Opt. Express, vol. 13, no. 19, pp. 7527-7534, 2005. 122 [22] S. J. Savory, "Digital coherent optical receivers: Algorithms and subsystems," IEEE J. Se!. Top. Quantum Electron., vol. 16, no. 5, pp. 1164-1179, 2010. [23] A. E. Willner, "Stable and Reconfigurable Optical Networks," in IEEE Lasers and Electro-Optics Society Annual Meeting 2007, Piscataway, NJ, USA, 2007, p. paper WFFl. [24] W. A. Atia and R. S. Bondurant, "Demonstration of return-to-zero signaling in both OOK and DPSK formats to improve receiver sensitivity in an optically pre-amplified receiver," presented at the LEOS'99, San Francisco, CA, 1999, Paper TuM3. [25] M. Hanna, H. Porte, J.-P. Goedgebuer, and W. T. Rhodes, "Performance assessment ofDPSK soliton transmission system," Electron. Lett., vol. 37, pp. 644-646, 2001. [26] M. Rohde, C. Caspar, N. Heimes, M. Konitzer, E.-J. Bachus, and N. Hanik, "Robustness of DPSK direct detection transmission format in standard fiber WDM systems," Electron. Lett., vol. 36, pp. 1483-1484, 2000. [27] A. H. Gnauck, G. Raybon, S. Chandrasekhar, J. Leuthold, C. Doerr, L. Stu!, A. Agarwal, S. Banerjee, D. Grosz, S. Hunsche, A. Kung, A. Marhelyuk, D. Maywar, M. Movassaghi, X. Liu, C. Xu, X. Wei, and D. M. Gill, "2.5 This (64 42.7 Gb/s) transmission over 40 100 km NZDSF using RZ-DPSK format and all-Raman-amplified spans," presented at the OFC'2002, Anaheim, CA, 2002, Postdeadline Paper FC-2. 123 [28] P. J. Winzer, A. H. Gnauck, C. R. Doerr, M. Magarini, and L. L. Buhl "Spectrally Efficient Long-Haul Optical Networking Using 112-Gb/s Polarization-Multiplexed 16- QAM," J. Lightwave Technol.28(4), 547-556 (2010). [29] A. H. Gnauck, G Charlet, P. Tran, P. J. Winzer, C. R. Doerr, J. C. Centanni, E. C. Burrows, T. Kawanishi, T. Sakamoto, and K. Higuma, "25.6-Tb/s WDM transmission of polarization-multiplexed RZ-DQPSK signals," J. Lightw. Technol., vol. 26, pp. 79-84, 2008. [30] A. H. Gnauck, P.J. Winzer, C. Dorrer, and S. Chandrasekhar, "Linear and nonlinear performance of 42.7-Gb/s single-polarization RZ-DQPSK format," Photonics Technology Letters, vol. 18, no. 7, pp. 883-885, Apr. 1, 2006. [31] C.R. S. Fludger, T. Duthel, D. van den Borne, C. Schulien, E.-D. Schmidt, T. Wuth, J Geyer, E. De Man, G.-D. Khoe, and H. de Waardt, "Coherent equalization and POLMUX-RZ-DQPSK for robust 100-GE transmission," J. Lightw. Technol., vol. 26, no. 1,pp.64-72,Jan.2008. [32] P. J. Winzer, "Modulation and multiplexing in optical communication systems," LEOS Newsletter Feb. 2009 [Online]. [33] R. Essiambre, G. J. Foschini, G. Kramer, and P. J. Winzer, "Capacity limits offiber optic networks," Phys. Rev. Lett., vol. 101, p. 163901, 2008. [34] R. Essiambre, G. Kramer, P. J. Winzer, G. J. Foschini, and B. Goebel, "Foundations for determining capacity limits of fibers," J. Lightw. Technol., 2009, accepted for publication. 124 [35] P. J. Winzer, "Advanced Optical Modulation Formats," Proceedings of the IEEE, Vol. 94, No. 5, May 2006. [36] D. A. Ackermann, J. E. Johnson, L. J. P. Ketelsen, L. E. Eng, P. A. Kiely, and T. G. B. Mason, "Telecommunication lasers," in Optical Fiber Telecommunications IV, I. Kaminow and T. Li, Eds. New York: Academic, 2002, pp. 587--065. [37] A. Ougazzaden, C. W. Lentz, T. G. B. Mason, K. G. Glogovsky, C. L. Reynolds, G. J. Przybylek, R. E. Leibenguth, T. L. Kercher, J. W. Boardman, M. T. Rader, J. M. Geary, F. S. Walters, J.M. F. L. J. Peticolas, S. N. G. Chu, A. Sirenka, R. J. Jurchenko, M. S. Hybertsen, L. J. P. Ketelsen, and G. Raybon, "40 Gb/s tandem electroabsorption modulator," in Proc. Optical Fiber Communication Conf. (OFC), 2001, Paper PD14. [38] J. Conradi, "Bandwidth-efficient modulation formats for digital fiber transmission systems," in Optical Fiber Telecommunications IV, I. Kaminow and T. Li, eds. New York: Academic, 2002, pp. 862-901. [39] A. H. Gnauck, S. K. Korotky, J. J. Veselka, J. Nagel, C. T. Kemmerer, W. J. Minford, and D. T. Moser, "Dispersion penalty reduction using an optical modulator with adjustable chirp," IEEE Photon. Technol. Lett., vol. 3, no. 10, pp. 916-918, Oct. 1991. [40] H. Kim and A. H. Gnauck, "Chirp characteristics of dual-drive Mach-Zehnder modulator with a finite DC extinction ratio," IEEE Photon. Technol. Lett., vol. 14, no. 3, pp. 298-300, Mar. 2002. 125 [41] P. J. Winzer. C. Dorrer. R.-J. Essiambre. and I. Kang. "Chirped return-to-zero modulation by imbalanced pulse carver driving signals," IEEE Photon. Technol. Lett., vol. 16, no. 5, pp. 1379-1381, May 2004. [42] G. Jacobsen, Noise in Digital Optical Transmission Systems. Norwood, MA: Artech House, 1994. [43] P. A. Humblet and M. Azizoglu, "On the bit error rate of lightwave systems with optical amplifiers," J. Llghtw. Technol., vol. 9, no. 11, pp. 1576-1582, Nov. 1991. [44] A. H. Gnauck and P. J. Winzer, "Optical phase-shift-keyed transmission," J. Lightwave Technol., vol. 23, no. 1, pp. 115-130, Jan. 2005. [45] P. J. Winzer, C. Dorrer, R.-J. Essiambre, and I. Kang, "Chirped return-to-zero modulation by imbalanced pulse carver driving signals," IEEE Photon. Technol. Lett., vol. 16, no. 5, pp. 1379-1381, May 2004. [46] R. J. Essiambre, G. Kramer, P. J. Winzer, G. J. Foschini and B. Goebel "Capacity limits of optical fiber networks," J. Lightwave Technol., vol. 28, no. 4, pp.662 -701 2010. [47] J. G. Proakis and M. Salehi, Digital Communications, 5th ed. New York: McGraw Hill, 2007. [48] C. E. Shannon, "A mathematical theory of communication," Bell Syst. Tech. J., vol. 27,pp. 379-423,623-656, 1948. [49] ITU-T Recommendation G975.l, Appendix I.9, 2004. 126 [50] T. Mizuochi et al., "Evolution and status of forward error correction," in Proc .Opt. Fiber Commun. Conf., Los Angeles, CA, 2012, Paper 0Tu2A.6. [51] F. Chang et al., "Forward error correction for 100 G transport networks," IEEE Commun. Mag., vol. 48, no. 3, pp. S48-S55, Mar. 2010. [52] G. D. Forney and G. Ungerboeck, "Modulation and coding for linear Gaussian channels," IEEE Trans. Inf. Theory, vol. 44, no. 6, pp. 2384-2415, Oct. 1998. [53] X. Liu et al., "Generation and FEC-decoding of a 231.5-Gb/s PDM-OFDM signal with 256-iterative-polar-modulation achieving 11.15-b/s/Hz intrachannel spectral efficiency and 800-km reach," in Proc. Opt. Fiber Commun. Conf., Los Angeles, CA, 2012, PDP5B.3. [54] J. Bromage, "Raman amplification for fiber communications systems," J. Lightw. Technol., vol. 22, no. 1. pp. 79-93, Jan. 2004. [55] E. Ip and J. M. Kahn, "Fiber impairment compensation using coherent detection and digital signal processing," J. Lightw. Technol., vol. 28, no. 4, pp. 502-519, Feb. 2010. [56] [89] P. J. Winzer, "Energy-efficient optical transport capacity scaling through spatial multiplexing," IEEE Photon. Technol. Lett., vol. 23, no. 13, pp. 851-853, Jul. 2011. [57] R.-J. Essiambre and R. W. Tkach, "Capacity trends and limits of optical communication networks," Proc. IEEE, vol. 100, no. 5, pp. 1035-1055, May 2012. [58] T. Morioka et al., "Enhancing optical communications with brand new fibers," IEEE Commun. Mag., vol. 50, no. 2, pp. s31-s42, Feb. 2012. 127 [59] P. J. Winzer, "Optical networking beyond WDM," IEEE Photon. J., vol. 4, no. 2, pp. 64 7-651, Apr. 2012. [60] Alcatel-Lucent 1830 Photonic Service Switch, Brochures and Data Sheets [Online]. Available: www.alcatel-lucent.com [61] J. Wang, J. Sun, X. Zhang, and D. Huang, "All-Optical Tunable Wavelength Conversion With Extinction Ratio Enhancement Using Periodically Poled Lithium Niobate Waveguides," J. Lightwave Technol. 26, 3137-3148 (2008). [62] X. Zhou and J. Yu, "200-Gb/s PDM-16QAM generation using a new synthesizing method," ECOC 2009, paper 10.3.5, 2009. [63] T. Sakamoto, A. Chiba, and T. Kawanishi, "50-Gb/s 16-QAM by a quad-parallel Mach-Zehnder modulator," ECOC 2007, paper PDP2.8, 2007. [64] P. J. Winzer, A. H. Gnauck, S. Chandrasekhar, S. Draving, J. Evangelista, and B. Zhu, "Generation and 1200-km transmission of 448-Gb/s ETDM 56-Gbaud PDM 16- QAM using a single l/Q modulator," in Proceedings of ECOC2010, paper PD2.2 (2010). [65] P. J. Winzer, "High-spectral-efficiency optical modulation formats," J. Lightwave Technol.30(24), 3824-3835 (2012). [66] Y. Koizumi, K. Toyoda, M. Yoshida, and M. Nakazawa, "1024 QAM (60 Gbit/s) single-carrier coherent optical transmission over 150 km," Opt. Express. 2011, 12508- 12514 (2012). 128 [67] P. J. Winzer. A. H. Gnauck. S. Chandrasekhar. S. Draving. J. Evangelista. and B. Zhu, "Generation and 1200-krn transmission of 448-Gb/s ETDM 56-Gbaud PDM 16- QAM using a single l/Q modulator," in Proceedings ofECOC2010, paper PD2.2 (2010). [68] T. Sakamoto, A. Chiba, and T. Kawanishi, "50-Gb/s 16-QAM by a quad-parallel Mach-Zehnder modulator," ECOC 2007, paper PDP2.8, 2007. [69] G.-W. Lu and T. Miyazaki, "Optical phase erasure based on FWM in HNLF enabling format conversion from 320-Gb/s RZ-DQPSK to160-Gb/s RZ-DPSK," Opt. Exp., vol. 17, pp. 13346-13353, 2009. [70] Mirco Scaffardi, Valeria Vercesi, Sergio Pinna, Antonella Bogoni, "All-Optical SOA-Assisted 40 Gbit/s DQPSK-to-OOK Format Conversion," Lecture Notes in Computer Science Volume 7715, 2013, pp 117-122. [71] J. Wang, Q. Sun and J. Sun "All-optical 40 Gb/s CSRZ-DPSK logic xor gate and format conversion using four-wave mixing," Opt. Express, vol. 17, no. 15, pp.12555 - 12563 2009. [72] G.-W. Lu, E. Tipsuwannakul, T. Miyazaki, C. Lundstrom, M. Karlsson and P.A. Andrekson "Format conversion of optical multilevel signals using FWM-based optical phase erasure," J. Lightw. Technol., vol. 29, no. 16, pp.2460 -2466 2011. [73] R. Slavik , J. Kakande , F. Parmigiani , P. Petropoulos and D. J. Richardson "All optical regeneration based on phase sensitive amplification," Proc. ConfLasers Electro- Opt., pp.1-22011. [74] R. Slavik et al, "All-optical phase and amplitude regenerator for next-generation telecommunications systems," Nature Photonics 4, 690--695 (2010). 129 [75] R. W. Tkach "Scaling Optical Communications for the Next Decade and Beyond," Bell Labs Tech. I., vol. 14, no. 4, pp.3 -10 2010. [76] P. Winzer "Beyond lOOG Ethernet," IEEE Commun. Mag., vol. 48, no. 7, pp.26 - 30 2010. [77] M. Matsumoto, "Performance Improvement of Phase-Shift-Keying Signal Transmission By Means of Optical Limiters Using Four-Wave Mixing in Fibers," IEEE J. Lightwave Technol. 23, 2696-2701 (2005). [78] M. Vasilyev, T. L. Lakoba, "All-optical multichannel 2R regeneration in a fiber based device," Opt. Lett. 30, 1458-1460 (2005). [79] S. Boscolo, S. K. Turitsyn, K. Blow, "All-optical passive 2R regeneration for N x 40 Gbit/s WDM transmission using NOLM and novel filtering technique," Opt. Commun. 217, 227-232 (2003). [80] P. Minzioni, V. Pusino, I. Cristiani, L. Marazzi, M. Martinelli, C. Langrock, M. M. Fejer, V. Degiorgio, "Optical phase conjugation in-phase-modulated transmission systems: experimental comparison of different nonlinearity-compensation methods," Opt. Express 18(17), 18119-18124 (2010). [81] S. L. Jansen, D. van den Borne, C. C. Monsalve, S. Spalter, P. M. Krummrich, G.D. Khoe, and H. de Waardt, "Reduction of Gordon-Mollenauer phase noise by mid-link spectral inversion," IEEE Photon. Technol. Lett. 17( 4), 923-925 (2005). [82] K. Cvecek, K. Sponsel, C. Stephan, G. Onishchukov, R. Ludwig, C. Schubert, B. Schmauss, and G. Leuchs, "Phase-preserving amplitude regeneration for a WDM RZ- 130 DPSK signal using a nonlinear amplifying loop mirror," Opt. Express 16(3), 1923-1928 (2008). [83] K. Croussore and G. Li "Amplitude regeneration of RZ-DPSK signals based on four-wave mixing in fibre," Electron. Lett., vol. 43, no. 3, pp.177 -178 2007 . [84] A. G. Striegler, M. Meissner, K. Cvecek, K. Sponsel, G. Leuchs, and B. Schmauss, "NOLM-Based RZ-DPSK Signal Regeneration," IEEE Photon. Technol. Lett. 17, 639- 641 (2005). [85] Y. Yu, W. Wu, X. Huang, B. Zou, S. Hu, X. Zhang, "Multi-channel all-optical RZ PSK amplitude regeneration based on the XPM effect in a single SOA," J. Lightw. Technol. 30, 3633-3639 (2012). [86] K. Inoue "Optical level equalisation based on gain saturation in optical parametric amplifiers," Electron. Lett., 2000, 36, pp. 1016-1017. [87] S. Radie, CJ. McKinstrie, R.M Jopson, J.C Centanni, and A.R Chraplyvy, "All optical regeneration in one- and two-pump parametric amplifiers using highly nonlinear optical fiber," IEEE Photonics Technol. Lett., 2003, 15, pp. 957-959. [88] P. Devgan, R. Tang, V.S. Grigoryan, and P. Kumar, "Multi-channel wavelength conversion of DPSK signals using FWM in highly nonlinear fiber without cross-gain modulation penalty," Conf. on Lasers and Electro-Optics (CLEO), 2005, paper CMQ3. [89] M. Matsumoto, "Performance improvement of DPSK signal transmission by means of optical limiters using FWM in fibers," IEEE J. Lightwave Technol., 2005, 23, pp. 2696-2701. 131 [90] M. Matsumoto. "Phase-preservation capability of all-optical amplitude regenerators using fiber nonlinearity," Conf. on Optical Fiber Communications (OFC), Anahain, CA, USA, 2006, paper JTHB.18. [91] M. Matsumoto, "Nonlinear phase noise reduction of DPSK signals by an all-optical amplitude limiter using FWM in a fiber," European Conf. on Optical Communications (ECOC) 2006, paper Tul.3.5. [92] N. J. Doran and D. Wood, "Picosecond soliton transmission using concatenated nonlinear optical loop-mirror intensity filters," J. Opt. Soc.Amer. B, vol. 12, no. 6, pp. 1117-1125, 1995. [93] F. Seguineau, B. Lavigne, D. Rouvillain, P. Brindel, L. Pierre, and 0. Leclerc, "Experimental demonstration of simple NOLM-based 2R-regenerator for 42.66 Gbit/s WDM long-haul transmissions," in Proc. OFC 2004, Los Angeles, CA, 2004, Paper WN4. [94] M. Meissner, K. Spouse!, K. Cvecek, A. Benz, S. Weisser, B. Schmauss, and G. Leuchs, "3.9-dB OSNR gain by an NOLM-based 2-R regenerator," IEEE Photon. Technol. Lett., vol. 16, no. 9, pp. 2105-2107, Sep. 2004. [95] T. Roethlingshoefer, T. Richter, C. Schubert, G. Onishchukov, B. Schmauss, and G. Leuchs, "All-optical phase preserving multilevel amplitude regeneration," Opt. Express 22(22), 27077-27085 (2014). [96] P. J. Winzer and H. Kim, "Degradations in balanced DPSK receivers," IEEE Photon. Technol. Lett., vol. 15, no. 9, pp. 1282-1284, Sep. 2003. 132 [97] Y. Yu, W. Wu, X. Huang, B. Zou, S. Hu, and X. Zhang, "Multi-channel all-optical RZ-PSK amplitude regeneration based on the XPM effect in a single SOA," J. Lightw. Technol., vol. 30, no. 23, pp. 3633-3639, Dec. 2012. [98] C. Porzi, A. Bogoni, and G. Contestabile, "Regeneration of DPSK signals in a saturated SOA," IEEE Photon. Technol. Lett., vol. 24, no. 18, pp. 1597-1599, Sep. 2012. [99] C. Porzi, A. Bogoni, and G. Contestabile, "Regenerative wavelength conversion of DPSK signals through FWM in an SOA," IEEE Photon. Technol. Lett., vol. 25, no. 2, pp. 175-178,Feb.2013. [100] K. Sato and H. Toba, "Reduction of mode partition noise by using semiconductor optical amplifiers," IEEE Se!. Topics Quantum Electron., vol. 7, no. 2, pp. 328-332, Mar./ Apr. 2001. [101] G. P. Agrawal and A. Olsson, "Self-phase modulation and spectral broadening of optical pulses in semiconductor laser amplifiers," IEEE J. Quantum Electron., vol. 25, no. 11, pp. 2297-2306, Nov. 1989. [102] B. Chen, and C. Xu, "Comparison of Cascaded i 2 l Wavelength Conversions in Quasi-Phase Matched (QPM) Waveguides," in SPIE Proceedings, vol. 5579, pp. 661 - 668, 2004. [103] H. Shankar, "Duobinary modulation for optical systems," White Paper, Inphi Corporation, (2004). [104] K. Y onenaga and S. Kuwano, "Dispersion-tolerant optical transmission system usmg duobinary transmitter and binary receiver," J. Lightwave Technol., vol. 15, pp.1530-1537 1997. 133 [105] M. Shtaif and A. H. Gnauck. "The relation between optical duobinary modulation and spectral efficiency in WDM systems," IEEE Photon. Technol. Lett. 11, 712-714 (1999). [106] K. Yonenaga, S. Kuwano, S. Narimatsu, and N. Shibata, "Optical duobinary transmission system with no receiver sensitivity degradation," Electron. Lett, vol. 31, no. 4, pp.302 -304 1995. [107] J. Ll , E. Tipsuwannakul , M. Karlsson and P. A. Andrekson "Low-complexity duobinary signaling and detectionfor sensitivity improvement in Nyquist-WDM coherent system," Opt. Fiber Commun. Conf/Nat. Fiber Opt. Eng. Conf, 2012. [108] C. Lam, "Google optical network," in proc. OFC/NFOEC, 2010, paper NWA3. [109] R. Essiambre, G. Kramer, P. J. Winzer, G. J. Foschini, and B. Goebel, "Capacity limits of optical fiber networks," J. Lightw. Technol., vol. 28, no. 4, pp.662-701, 2010. [110] L. G. Kazovsky, G. Kalogerakis, and W.-T. Shaw, "Homodyne phase-shift keying systems: Past challenges and future opportunities," J. Lightw. Technol., vol. 24, no. 12, pp.4876-4884,2006. [111] R. Noe, "Phase noise-tolerant synchronous QPSK/BPSK baseband-type intradyne receiver concept with feedforward carrier recovery," J. Lightw. Technol., vol.23, no.2, pp.802-808, 2005. [112] R. Freund, L. Molle, F. Raub, C. Caspar, M. Karkri, and C. Weber, "Triple (S/C/L)-band WDM transmission using erbium-doped fibre amplifiers," in Proc. ECOC, 2005, paper Mo.4.2.3. 134 [113] A. Leven, N. Kaneda, and Y. K.Chen, "A real-time CMA-based lOGb/s polarization demultiplexing coherent receiver implemented in an FPGA," in Proc. OFC/NFOEC, 2008, paper Otu02. [114] D.-S. Ly-Gagnon, S. Tsukamoto, K. Katoh, and K. Kikuchi, "Coherent detection of optical quadrature phase-shift keying signals with carrier phase estimation," J. Lightw. Technol., vol. 24, no. 1, pp. 12-21, 2006. [115] S. J. Savory, G. Gavioli, R. I. Killey, and P. Bayvel, "Transmission of 42.8Gbit/s polarization multiplexed NRZ-QPSK over 6400km of standard fiber with no optical dispersion compensation," in Proc. OFC, 2007, paper OtuAl. [116] E. Ip, A. P. T. Lau, D. J. F. Barros, and J. M. Kahn, "Coherent detection in optical fiber systems," Opt. Express, vol. 16, no. 2, pp. 753-791, 2008. [117] M. G. Taylor, "Coherent detection method using DSP for demodulation of signal and subsequent equalization of propagation impairments," IEEE Photon. Technol. Lett., vol. 16,no.2,pp.674--676,2004. [118] H. Sun, K.-T.Wu, and K. Roberts, "Real-time measurements of a 40 Gb/s coherent system," Opt. Express, vol. 16, no. 2, pp. 873-879, 2008. [119] J. Yu, M.-F. Huang, S. Zhang, L. Xu, Y. Inada, T. Ogata, and Y. Aoki, "Transmission of 42.8-Gb/s polarization multiplexed RZ-QPSK DWDM signals over 3900 km with 12.5-Ghz channel spacing and coherent detection," in Proc. OFC/NFOEC, 2010, paper OTuD4. 135 [120] C. Fludger, T, Duthel, D, van den Borne, C. Schulien, K-D, Schmidt, T,Wuth, J, Geyer, E, DeMan, K, Giok-Djan, and H, deWaardt, "Coherent equalization and POLMUX-RZ-DQPSK for robust 100-GE transmission," l Lightw, TechnoL, voL 26, no, 1, pp, 64-72, 2008, [121] P, l Winzer, "High-spectral-efficiency optical modulation formats," l Lightwave TechnoL30(24), 3824-3835 (2012), [122] R, Slavik, J, Kakande , F Parmigiani , P, Petropoulos and D, J, Richardson "All optical regeneration basedon phase sensitive amplification," Proc, ConfLasers Electro- Opt,, pp,l -2 201 L [123] G,-W, Lu and T, Miyazaki, "Optical phase erasure based on FWM in HNLF enabling format conversion from 320-Gb/s RZ-DQPSK to160-Gb/s RZ-DPSK," Opt Exp,, voL 17, pp, 13346-13353, 2009, [124] G,-W, Lu, E, Tipsuwannakul, T, Miyazaki, C. Lundstrom, M, Karlsson and P, A, Andrekson "Format conversion of optical multilevel signals using FWM-based optical phase erasure," J, Lightw, TechnoL, voL 29, no, 16, pp,2460 -2466 201 L [125] C. Porzi, A, Bogoni, G, Contestabile, "Regenerative wavelength conversion of DPSK signals through FWM in an SOA," IEEE Photon, TechnoL Lett 25, 175-178 (2013), [126] K, Sato, H, Toba, "Reduction of mode partition noise by using semiconductor optical amplifiers," IEEE SeL Topics Quantum Electron, 7, 328-332 (2001), 136 [127] G. P. Agrawal. A. Olsson. "Self-phase modulation and spectral broadening of optical pulses in semiconductor laser amplifiers," IEEE J. Quantum Electron. 25, 2297- 2306 (1989 ). [128] M. Matsumoto, "Performance Improvement of Phase-Shift-Keying Signal Transmission By Means of Optical Limiters Using Four-Wave Mixing in Fibers," IEEE J. Lightwave Technol. 23, 2696-2701 (2005). [129] P. J. Winzer and R.-J. Essiambre, "Advanced Modulation Formats for High capacity Optical Transport Networks," J. Lightwave Technol. 24(12), 4711-4728 (2006). [130] M. Seimetz, High-Order Modulation for Optical Fiber Transmission, Springer Series in Optical Sciences, Springer, (2009). [131] J. P. Gordon and L. F. Mollenauer, "Phase Noise in Photonic Communications Systems using Linear Amplifiers," Opt. Lett. 15(23), 1351-1353 (1990). [132] H. Kim, "Cross-Phase-Modulation-Induced Nonlinear Phase Noise m WDM Direct-Detection DPSK Systems," J. Lightwave Technol. 21(8), 1770-1774 (2003). [133] M. Gao, T. Inoue, T. Kurosu, and S. Namiki, "Sideband-assisted phase sensitive amplifiers with high phase sensitivity for efficient phase regeneration," in Proc. OFC2012, paper OW3C.5 (2012). [134] B. Stiller, G. Onishchukov, B. Schmauss, and G. Leuchs, "Phase regeneration of a star-8QAM signal in a phase-sensitive amplifier with conjugated pumps," Opt. Express 22, 1028-1035 (2014). [135] M. Hierold, T. Roethlingshoefer, K. Sponsel, G. Onishchukov, B. Schmauss, and G. Leuchs, "Multilevel Phase preserving Amplitude Regeneration using a Single 137 Nonlinear Amplifying Loop Mirror," IEEE Photon. Technol. Lett. 23(14), 1007-1009 (2011). [136] T. Roethlingshoefer, T. Richter, C. Schubert, G. Onishchukov, B. Schmauss, and G. Leuchs, "All-optical phase preserving multilevel amplitude regeneration," Opt. Express 22(22), 27077-27085 (2014). [137] Z. Bakhtiari, J. Wang, X. Wu, J. Yang, S. Nuccio, R. Hellwarth, A. Willner, "Demonstration of 10-40-Gbaud baud-rate-tunable optical generation of 16-QAM from a QPSK signal using a variable DGD element," in Proc. CLE0201 l, paper CThXS (2011). [138] C. Porzi, G. Serafino, A. Bogoni, G. Contestabile, "Phase-Presrving Amplitude Noise Compression of 40 Gb/s DPSK Signals in a Single SOA," J. Lightwave Technol. 32(10), 1966-1972 (2014). [139] M. Sorokina, "ultilevel amplitude regeneration of 256-symbol constellation," in Proc. CLE02011, paper STu2J.2 (2014). [140] P. J. Winzer and R.-J. Essiambre, "Advanced Modulation Formats for HighCapacity Optical Transport Networks," Lightwave Technology, Journal of, vol. 24, pp. 4711-4728, 2006. [141] A. Yariv and P. Yeh, "Optical Waves in Crystals: Propagation and Control of Laser Radiation," New York: Wiley-Interscience, 1984. [142] A. H. Gnauck and P. J. Winzer, "Optical phase-shift-keyed transmission," Lightwave Technology, Journal of, vol. 23, pp. 115-130, 2005. [143] J. R. Andrews, "RZ vs. NRZ," 2001. 138 [144] J. D. Downie and A. B. Ruffin. "Analysis of signal distortion and crosstalk penalties induced by optical filters in optical networks," Llghtwave Technology, Journal of, vol. 21, pp. 1876-1886, 2003. 139

Abstract (if available)

Abstract In the next generation of optical networks, high speed data rates of 100Gb/s per channel or more will be required. In order to include more channels in a single fiber in wavelength division multiplexing (WDM) systems, the channel spacing must be decreased from 200GHz to 50GHz or even smaller. These dense high bit rate WDM systems suffer more severely from the effect of linear and nonlinear degradation in fiber transmission. Phase-modulated format-based WDM systems that have more tolerance for degrading effects, particularly nonlinearity-based degradation have been explored. With growing demand for transmission capacity along with the desire to lower the cost per information bit, spectrally-efficient multilevel modulation format signals such as quadrature amplitude modulation (QAM) have become popular. Higher-order data modulation format signals are quite important to optical communications due to their high spectral efficiency, low electrical baud rate and increased tolerance to fiber-based impairments. Specifically, there is interest in generating quadrature-amplitude-modulation (QAM) signals, and researchers have demonstrated up to 1024-QAM. A laudable goal is the generation of high-order QAM in a tunable fashion such that variable bit rates can be accommodated using optical methods. ❧ In this dissertation, we demonstrate two different techniques for generation of QAM signals. The first method is a polarization-based technique for high speed, tunable QAM signal generation that implements the amplitude control outside the integrated device using a polarizer. The second method is a nonlinearity-based technique that multiplexes initial lower-level modulation format signals at different frequencies into a QAM signal. We also develop fully optical modules to apply various signal processing operations to QAM signals, e.g. demultplexing and information extraction as techniques to avoid coherent receivers in the middle of optical networks. The optical demultiplexing module is based on the phase-sensitive amplification concept, and can demultiplex a QPSK signal into two BPSK sub-channels. We also develop a fully optical module that provides logic/arithmetic relations between symbols carried by a QAM signal. ❧ A disadvantage of higher-order QAM signals is higher sensitivity to noise accumulation, especially in long-haul transmission systems. Amplitude noise not only reduces signal quality but may also be converted into nonlinear phase noise in a transmission line due to the Gordon–Mollenauer effect. Cross-phase modulation in wavelength-division multiplexing (WDM) systems is another cause of amplitude noise to nonlinear phase noise conversion. All-optical regenerators are expected to extend the maximum reach of high-speed transmission systems by eliminating accumulated signal impairments in transmission systems without the need for optical/electronic/optical (O/E/O) conversion. Current phase regeneration schemes are relatively complex and are limited to lower-order modulation formats. Thus, developing all-optical tunable phase-preserving multilevel amplitude regenerator modules with bit-rate transparency is desirable. ❧ In this dissertation, we demonstrate all-optical multilevel amplitude regeneration for various types of star-QAM and also square-16QAM signals using a technique that we call coherent polarization mixing. We describe two different schemes for the proposed polarization-based multilevel amplitude regeneration technique. The schemes provide a wide range of tunability and scalability, and have a simpler configuration compared to previous methods.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

High-speed and reconfigurable all-optical signal processing for phase and amplitude modulated signals

PDF

Reconfigurable high‐speed optical signal processing and high‐capacity optical transmitter

PDF

Reconfigurable optical signal processing for efficient spectrum utilization in high-speed optical communication systems

PDF

Optical wave mixing for tunable delays and high‐speed signal processing

PDF

Reconfigurable and flexible high-speed optical signal processing and spectrally shared optical subsystems

PDF

Applications of all optical signal processing for advanced optical modulation formats

PDF

Optical signal processing for enabling high-speed, highly spectrally efficient and high capacity optical systems

PDF

Reconfigurable high-speed processing and noise mitigation of optical data

PDF

Reconfigurable high speed optical signal processing for optical communication and modulation format manipulation

PDF

Optical signal processing for high-speed, reconfigurable fiber optic networks

PDF

Integrated silicon waveguides and specialty optical fibers for optical communications system applications

PDF

Nonlinear optical signal processing for high-speed, spectrally efficient fiber optic systems and networks

PDF

Detection and optical signal processing using phase based optical modulation formats

PDF

Advanced modulation, detection, and monitoring techniques for optical communication systems

PDF

Microresonator-based Kerr frequency comb for high-speed optical communications and signal processing

PDF

Orbital angular momentum based spatially multiplexed optical and millimeter-wave communications

PDF

Spectrally efficient waveforms for coherent optical fiber transmission

PDF

High-capacity optical communications using structured light in random and disturbed media

PDF

High capacity optical and RF/mm-wave communications utilizing orbital angular momentum multiplexing

PDF

Optical communications, optical ranging, and optical signal processing using tunable and reconfigurable technologies

Asset Metadata

Creator Bakhtiari, Zahra (author)

Core Title All-optical signal processing toward reconfigurable optical networks

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Electrical Engineering

Publication Date 01/27/2016

Defense Date 03/10/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag coherent systems,fiber optics communications,fiber optics links and subsystems,nonlinear optics,OAI-PMH Harvest,optical signal processing

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Sawchuk, Alexander A. (Sandy) (committee chair), Armani, Andrea (committee member), Haas, Stephan (committee member), Jenkins, B. Keith (committee member), Molisch, Andreas F. (committee member), Willner, Alan E. (committee member)

Creator Email mehrnaz08@gmail.com,zbakhtia@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-611390

Unique identifier UC11303776

Identifier etd-BakhtiariZ-3730.pdf (filename),usctheses-c3-611390 (legacy record id)

Legacy Identifier etd-BakhtiariZ-3730.pdf

Dmrecord 611390

Document Type Dissertation

Format application/pdf (imt)

Rights Bakhtiari, Zahra

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA